Search

Mechanistic models have been used for centuries to describe complex interconnected processes, including biological ones. As the scope of these models has widened, so have their computational demands. This complexity can limit their suitability when running many simulations or when real-time results are required. Surrogate machine learning (ML) models can be used to approximate the behaviour of complex mechanistic models, and once built, their computational demands are several orders of magnitude lower. This paper provides an overview of the relevant literature, both from an applicability and a theoretical perspective. For the latter, the paper focuses on the design and training of the underlying ML models. Application-wise, we show how ML surrogates have been used to approximate different mechanistic models. We present a perspective on how these approaches can be applied to models representing biological processes with potential industrial applications (e.g., metabolism and whole-cell modelling) and show why surrogate ML models may hold the key to making the simulation of complex biological systems possible using a typical desktop computer.

Reliable and robust simulation of individual patients using patient-specific models (PSMs) is one of the next frontiers for modeling and simulation (M&S) in healthcare. PSMs, which form the basis of digital twins, can be employed as clinical tools to, for example, assess disease state, predict response to therapy, or optimize therapy. They may also be used to construct virtual cohorts of patients, for in silico evaluation of medical product safety and/or performance. Methods and frameworks have recently been proposed for evaluating the credibility of M&S in healthcare applications. However, such efforts have generally been motivated by models of medical devices or generic patient models; how best to evaluate the credibility of PSMs has largely been unexplored. The aim of this paper is to understand and demonstrate the credibility assessment process for PSMs using patient-specific cardiac electrophysiological (EP) modeling as an exemplar. We first review approaches used to generate cardiac PSMs and consider how verification, validation, and uncertainty quantification (VVUQ) apply to cardiac PSMs. Next, we execute two simulation studies using a publicly available virtual cohort of 24 patient-specific ventricular models, the first a multi-patient verification study, the second investigating the impact of uncertainty in personalized and non-personalized inputs in a virtual cohort. We then use the findings from our analyses to identify how important characteristics of PSMs can be considered when assessing credibility with the approach of the ASME V&V40 Standard, accounting for PSM concepts such as inter- and intra-user variability, multi-patient and “every-patient” error estimation, uncertainty quantification in personalized vs non-personalized inputs, clinical validation, and others. The results of this paper will be useful to developers of cardiac and other medical image based PSMs, when assessing PSM credibility.

The design of efficient combination therapies is a difficult key challenge in the treatment of complex diseases such as cancers. The large heterogeneity of cancers and the large number of available drugs renders exhaustive in vivo or even in vitro investigation of possible treatments impractical. In recent years, sophisticated mechanistic, ordinary differential equation-based pathways models that can predict treatment responses at a molecular level have been developed. However, surprisingly little effort has been put into leveraging these models to find novel therapies. In this paper we use for the first time, to our knowledge, a large-scale state-of-the-art pan-cancer signaling pathway model to identify candidates for novel combination therapies to treat individual cancer cell lines from various tissues (e.g., minimizing proliferation while keeping dosage low to avoid adverse side effects) and populations of heterogeneous cancer cell lines (e.g., minimizing the maximum or average proliferation across the cell lines while keeping dosage low). We also show how our method can be used to optimize the drug combinations used in sequential treatment plans—that is, optimized sequences of potentially different drug combinations—providing additional benefits. In order to solve the treatment optimization problems, we combine the Covariance Matrix Adaptation Evolution Strategy (CMA-ES) algorithm with a significantly more scalable sampling scheme for truncated Gaussian distributions, based on a Hamiltonian Monte-Carlo method. These optimization techniques are independent of the signaling pathway model, and can thus be adapted to find treatment candidates for other complex diseases than cancers as well, as long as a suitable predictive model is available., Author summary Combination therapies are a promising approach to counter complex diseases such as cancers. Two key difficulties in the design of effective cancer combination therapies are the large number of available drugs and the heterogeneity of cancers which render exhaustive laboratory studies impractical. In recent years, sophisticated signaling pathway models that can predict responses to combination treatments at a molecular level have been developed. This motivates the question of how one can leverage mechanistic models to identify candidates for novel combination treatments. In this paper we propose a combination treatment optimization framework which employs a large-scale pan-cancer pathway model. We formulate treatment optimization problems for single cell lines and heterogeneous populations of cancer cells. We further investigate sequential treatment plans and combine an existing evolutionary algorithm with an efficient Hamiltonian Monte-Carlo based sampling scheme. During extensive simulation studies our approach identified combination therapies which are predicted to be more effective than conventional treatments. We hope that one day in silico experiments will be used to identify a small set of promising treatment candidates which can then form a starting point for laboratory studies, allowing for an efficient use of limited resources and accelerated discovery of effective therapies.

In their Commentary paper, Villaverde and Massonis (On testing structural identifiability by a simple scaling method: relying on scaling symmetries can be misleading) have commented on our paper in which we proposed a simple scaling method to test structural identifiability. Our scaling invariance method (SIM) tests for scaling symmetries only, and Villaverde and Massonis correctly show the SIM may fail to detect identifiability problems when a model has other types of symmetries. We agree with the limitations raised by these authors but, also, we emphasize that the method is still valuable for its applicability to a wide variety of models, its simplicity, and even as a tool to introduce the problem of identifiability to investigators with little training in mathematics.

Recurrent neural networks with memory and attention mechanisms are widely used in natural language processing because they can capture short and long term sequential information for diverse tasks. We propose an integrated deep learning model for microbial DNA sequence data, which exploits convolutional neural networks, recurrent neural networks, and attention mechanisms to predict taxonomic classifications and sample-associated attributes, such as the relationship between the microbiome and host phenotype, on the read/sequence level. In this paper, we develop this novel deep learning approach and evaluate its application to amplicon sequences. We apply our approach to short DNA reads and full sequences of 16S ribosomal RNA (rRNA) marker genes, which identify the heterogeneity of a microbial community sample. We demonstrate that our implementation of a novel attention-based deep network architecture, Read2Pheno, achieves read-level phenotypic prediction. Training Read2Pheno models will encode sequences (reads) into dense, meaningful representations: learned embedded vectors output from the intermediate layer of the network model, which can provide biological insight when visualized. The attention layer of Read2Pheno models can also automatically identify nucleotide regions in reads/sequences which are particularly informative for classification. As such, this novel approach can avoid pre/post-processing and manual interpretation required with conventional approaches to microbiome sequence classification. We further show, as proof-of-concept, that aggregating read-level information can robustly predict microbial community properties, host phenotype, and taxonomic classification, with performance at least comparable to conventional approaches. An implementation of the attention-based deep learning network is available at https://github.com/EESI/sequence_attention (a python package) and https://github.com/EESI/seq2att (a command line tool)., Author summary Microbiomes are communities of microscopic organisms found everywhere, including on and in the human body. For example, the gut microbiome plays an important role in digestion, and changes in composition are associated with changes in health or disease, e.g., inflammatory bowel disease (IBD). Today, microbiome composition is often obtained from high-throughput sequencing, which generates many short DNA reads from multiple organisms in a sample. In this paper, we present a novel deep learning framework, Read2Pheno, to predict phenotype from all the reads in a set of biological samples. An attention mechanism allows visualization of specific subregions (sets of bases) which are important in classifying the reads according to phenotype or taxon labels. We evaluate the framework on sequencing data for 16S rRNA genes, genetic markers used to identify microbial taxonomy. We show that Read2Pheno performs comparably as conventional methods on three distinct data sets from the American Gut Project, IBD patients and controls, and a comprehensive taxonomic database. Moreover, Read2Pheno results can be readily interpreted—e.g., to identify regions of the 16S rRNA gene to target for PCR diagnostics—without additional pre/post-processing steps that can introduce complexity and error.

Mathematical models are powerful tools in HIV epidemiology, producing quantitative projections of key indicators such as HIV incidence and prevalence. In order to improve the accuracy of predictions, such models need to incorporate a number of behavioural and biological heterogeneities, especially those related to the sexual network within which HIV transmission occurs. An individual-based model, which explicitly models sexual partnerships, is thus often the most natural type of model to choose. In this paper we present PopART-IBM, a computationally efficient individual-based model capable of simulating 50 years of an HIV epidemic in a large, high-prevalence community in under a minute. We show how the model calibrates within a Bayesian inference framework to detailed age- and sex-stratified data from multiple sources on HIV prevalence, awareness of HIV status, ART status, and viral suppression for an HPTN 071 (PopART) study community in Zambia, and present future projections of HIV prevalence and incidence for this community in the absence of trial intervention., Author summary In this paper we present PopART-IBM, an individual-based model used to simulate HIV transmission in communities in high prevalence settings. We show that PopART-IBM can simulate transmission over a span of decades in a large community in less than a minute. This computational efficiency allows us to calibrate the model within an inference framework, and we show an illustrative example of calibration using an adaptive population Monte Carlo Approximate Bayesian Computation algorithm for a community in Zambia that was part of the HPTN-071 (PopART) trial. We compare the detailed model output to real-world data collected during the trial from this community. Finally, we project how the HIV epidemic would have changed over time in this community if no intervention from the trial had occurred.

Stochastic models of biomolecular reaction networks are commonly employed in systems and synthetic biology to study the effects of stochastic fluctuations emanating from reactions involving species with low copy-numbers. For such models, the Kolmogorov’s forward equation is called the chemical master equation (CME), and it is a fundamental system of linear ordinary differential equations (ODEs) that describes the evolution of the probability distribution of the random state-vector representing the copy-numbers of all the reacting species. The size of this system is given by the number of states that are accessible by the chemical system, and for most examples of interest this number is either very large or infinite. Moreover, approximations that reduce the size of the system by retaining only a finite number of important chemical states (e.g. those with non-negligible probability) result in high-dimensional ODE systems, even when the number of reacting species is small. Consequently, accurate numerical solution of the CME is very challenging, despite the linear nature of the underlying ODEs. One often resorts to estimating the solutions via computationally intensive stochastic simulations. The goal of the present paper is to develop a novel deep-learning approach for computing solution statistics of high-dimensional CMEs by reformulating the stochastic dynamics using Kolmogorov’s backward equation. The proposed method leverages superior approximation properties of Deep Neural Networks (DNNs) to reliably estimate expectations under the CME solution for several user-defined functions of the state-vector. This method is algorithmically based on reinforcement learning and it only requires a moderate number of stochastic simulations (in comparison to typical simulation-based approaches) to train the “policy function”. This allows not just the numerical approximation of various expectations for the CME solution but also of its sensitivities with respect to all the reaction network parameters (e.g. rate constants). We provide four examples to illustrate our methodology and provide several directions for future research., Author summary We develop a deep learning framework for estimating solutions of the chemical master equation (CME) that is fundamental to stochastic analysis of reaction networks. The CME is a system of ordinary differential equations that describes the time-evolution of the probability density of the random state-vector, and owing to an inherent curse of dimensionality, directly solving the CME is generally impractical with existing approaches. Moreover, the commonly employed simulation-based approaches for estimating CME solutions often require an exorbitant amount of computational time, even for moderately-sized networks. To counter these issues, we develop a deep reinforcement learning based method, called DeepCME, in this paper. DeepCME not only estimates function expectations based on the CME solution, but it also solves the more challenging problem of estimating their sensitivities with respect to all the model parameters. We illustrate our approach with four carefully chosen reaction network examples with varying sizes. Our results demonstrate that DeepCME reliably estimates the expectations of interest, along with all the parametric sensitivities, at a fraction of the computational cost of simulation-based estimators. We present many directions for future research and suggest further improvements to DeepCME that can greatly enhance its accuracy and applicability.

High-throughput data such as metabolomics, genomics, transcriptomics, and proteomics have become familiar data types within the “-omics” family. For this work, we focus on subsets that interact with one another and represent these “pathways” as graphs. Observed pathways often have disjoint components, i.e., nodes or sets of nodes (metabolites, etc.) not connected to any other within the pathway, which notably lessens testing power. In this paper we propose the Pathway Integrated Regression-based Kernel Association Test (PaIRKAT), a new kernel machine regression method for incorporating known pathway information into the semi-parametric kernel regression framework. This work extends previous kernel machine approaches. This paper also contributes an application of a graph kernel regularization method for overcoming disconnected pathways. By incorporating a regularized or “smoothed” graph into a score test, PaIRKAT can provide more powerful tests for associations between biological pathways and phenotypes of interest and will be helpful in identifying novel pathways for targeted clinical research. We evaluate this method through several simulation studies and an application to real metabolomics data from the COPDGene study. Our simulation studies illustrate the robustness of this method to incorrect and incomplete pathway knowledge, and the real data analysis shows meaningful improvements of testing power in pathways. PaIRKAT was developed for application to metabolomic pathway data, but the techniques are easily generalizable to other data sources with a graph-like structure., Author summary PaIRKAT is a tool for improving testing power on high dimensional data by including graph topography in the kernel machine regression setting. Studies on high-dimensional data can struggle to include the complex relationships between variables. The semi-parametric kernel machine regression model is a powerful tool for capturing these types of relationships. They provide a framework for testing for relationships between outcomes of interest and high dimensional data such as metabolomic, genomic, or proteomic pathways. Our paper proposes a kernel machine method for including known biological connections between high dimensional variables by representing them as edges of ‘graphs’ or ‘networks.’ It is common for nodes (e.g., metabolites) to be disconnected from all others within the graph, which leads to meaningful decreases in testing power when graph information is ignored. We include a graph regularization or ‘smoothing’ approach for managing this issue. We demonstrate the benefits of this approach through simulation studies and an application to the metabolomic data from the COPDGene study.

The ever increasing applications of bioinformatics in providing effective interpretation of large and complex biological data require expertise in the use of sophisticated computational tools and advanced statistical tests, skills that are mostly lacking in the Sudanese research community. This can be attributed to paucity in the development and promotion of bioinformatics, lack of senior bioinformaticians, and the general status quo of inadequate research funding in Sudan. In this paper, we describe the challenges that have encountered the development of bioinformatics as a discipline in Sudan. Additionally, we highlight on specific actions that may help develop and promote its education and training. The paper takes the National University Biomedical Research Institute (NUBRI) as an example of an institute that has tackled many of these challenges and strives to drive powerful efforts in the development of bioinformatics in the country., Author summary Bioinformatics is gaining recognition globally and in Sudan as an important subdiscipline of biological sciences, one that enables researchers to efficiently interpret complex biological data. A limited number of Sudanese academic institutions have acknowledged this field despite its increasingly recognized importance. The development of bioinformatics in the country requires interdisciplinary collaborations involving experts in life sciences, research methodology, healthcare, computer, and data sciences. This can be achieved through designing educational programs and workshops alongside proposing and establishing effective collaborative research projects. In this context, we comprehensively discussed the present state of bioinformatics in Sudan, the challenges faced, as well as the efforts exerted by academic institutions including NUBRI, to upgrade infrastructure and establish local and international collaborations.

In their Commentary paper, Villaverde and Massonis (On testing structural identifiability by a simple scaling method: relying on scaling symmetries can be misleading) have commented on our paper in which we proposed a simple scaling method to test structural identifiability. Our scaling invariance method (SIM) tests for scaling symmetries only, and Villaverde and Massonis correctly show the SIM may fail to detect identifiability problems when a model has other types of symmetries. We agree with the limitations raised by these authors but, also, we emphasize that the method is still valuable for its applicability to a wide variety of models, its simplicity, and even as a tool to introduce the problem of identifiability to investigators with little training in mathematics.

Read-depths (RDs) are frequently used in identifying structural variants (SVs) from sequencing data. For existing RD-based SV callers, it is difficult for them to determine breakpoints in single-nucleotide resolution due to the noisiness of RD data and the bin-based calculation. In this paper, we propose to use the deep segmentation model UNet to learn base-wise RD patterns surrounding breakpoints of known SVs. We integrate model predictions with an RD-based SV caller to enhance breakpoints in single-nucleotide resolution. We show that UNet can be trained with a small amount of data and can be applied both in-sample and cross-sample. An enhancement pipeline named RDBKE significantly increases the number of SVs with more precise breakpoints on simulated and real data. The source code of RDBKE is freely available at https://github.com/yaozhong/deepIntraSV., Author summary In this paper, we used the deep segmentation model UNet to alleviate the bin size limitation of RD-based SV callers. UNet was initially proposed for image data. Here, we demonstrated that the UNet model could also be applied for one-dimensional genomic data. We formalized the breakpoint prediction as a segmentation task and inferred breakpoints in single-nucleotide resolution from predicted label marks. Through a set of experiments on both simulated and real WGS data, we demonstrated that the UNet model could be trained with a small amount of data, and an enhancement pipeline called RDBKE significantly increased the number of SVs with more precise breakpoints.

The ubiquitous availability of genome sequencing data explains the popularity of machine learning-based methods for the prediction of protein properties from their amino acid sequences. Over the years, while revising our own work, reading submitted manuscripts as well as published papers, we have noticed several recurring issues, which make some reported findings hard to understand and replicate. We suspect this may be due to biologists being unfamiliar with machine learning methodology, or conversely, machine learning experts may miss some of the knowledge needed to correctly apply their methods to proteins. Here, we aim to bridge this gap for developers of such methods. The most striking issues are linked to a lack of clarity: how were annotations of interest obtained; which benchmark metrics were used; how are positives and negatives defined. Others relate to a lack of rigor: If you sneak in structural information, your method is not sequence-based; if you compare your own model to “state-of-the-art,” take the best methods; if you want to conclude that some method is better than another, obtain a significance estimate to support this claim. These, and other issues, we will cover in detail. These points may have seemed obvious to the authors during writing; however, they are not always clear-cut to the readers. We also expect many of these tips to hold for other machine learning-based applications in biology. Therefore, many computational biologists who develop methods in this particular subject will benefit from a concise overview of what to avoid and what to do instead.

Single cell RNA sequencing (scRNA-seq) enables researchers to characterize transcriptomic profiles at the single-cell resolution with increasingly high throughput. Clustering is a crucial step in single cell analysis. Clustering analysis of transcriptome profiled by scRNA-seq can reveal the heterogeneity and diversity of cells. However, single cell study still remains great challenges due to its high noise and dimension. Subspace clustering aims at discovering the intrinsic structure of data in unsupervised fashion. In this paper, we propose a deep sparse subspace clustering method scDSSC combining noise reduction and dimensionality reduction for scRNA-seq data, which simultaneously learns feature representation and clustering via explicit modelling of scRNA-seq data generation. Experiments on a variety of scRNA-seq datasets from thousands to tens of thousands of cells have shown that scDSSC can significantly improve clustering performance and facilitate the interpretability of clustering and downstream analysis. Compared to some popular scRNA-deq analysis methods, scDSSC outperformed state-of-the-art methods under various clustering performance metrics.

Synaptic dysfunction is associated with several brain disorders, including Alzheimer’s disease, Parkinson’s disease (PD) and obsessive compulsive disorder (OCD). Utilizing synaptic plasticity, brain stimulation is capable of reshaping synaptic connectivity. This may pave the way for novel therapies that specifically counteract pathological synaptic connectivity. For instance, in PD, novel multichannel coordinated reset stimulation (CRS) was designed to counteract neuronal synchrony and down-regulate pathological synaptic connectivity. CRS was shown to entail long-lasting therapeutic aftereffects in PD patients and related animal models. This is in marked contrast to conventional deep brain stimulation (DBS) therapy, where PD symptoms return shortly after stimulation ceases. In the present paper, we study synaptic reshaping by periodic multichannel stimulation (PMCS) in networks of leaky integrate-and-fire (LIF) neurons with spike-timing-dependent plasticity (STDP). During PMCS, phase-shifted periodic stimulus trains are delivered to segregated neuronal subpopulations. Harnessing STDP, PMCS leads to changes of the synaptic network structure. We found that the PMCS-induced changes of the network structure depend on both the phase lags between stimuli and the shape of individual stimuli. Single-pulse stimuli and burst stimuli with low intraburst frequency down-regulate synapses between neurons receiving stimuli simultaneously. In contrast, burst stimuli with high intraburst frequency up-regulate these synapses. We derive theoretical approximations of the stimulation-induced network structure. This enables us to formulate stimulation strategies for inducing a variety of network structures. Our results provide testable hypotheses for future pre-clinical and clinical studies and suggest that periodic multichannel stimulation may be suitable for reshaping plastic neuronal networks to counteract pathological synaptic connectivity. Furthermore, we provide novel insight on how the stimulus type may affect the long-lasting outcome of conventional DBS. This may strongly impact parameter adjustment procedures for clinical DBS, which, so far, primarily focused on acute effects of stimulation.

Natural products are chemical compounds that form the basis of many therapeutics used in the pharmaceutical industry. In microbes, natural products are synthesized by groups of colocalized genes called biosynthetic gene clusters (BGCs). With advances in high-throughput sequencing, there has been an increase of complete microbial isolate genomes and metagenomes, from which a vast number of BGCs are undiscovered. Here, we introduce a self-supervised learning approach designed to identify and characterize BGCs from such data. To do this, we represent BGCs as chains of functional protein domains and train a masked language model on these domains. We assess the ability of our approach to detect BGCs and characterize BGC properties in bacterial genomes. We also demonstrate that our model can learn meaningful representations of BGCs and their constituent domains, detect BGCs in microbial genomes, and predict BGC product classes. These results highlight self-supervised neural networks as a promising framework for improving BGC prediction and classification.Author summaryBiosynthetic gene clusters (BGCs) encode for natural products of diverse chemical structures and function, but they are often difficult to discover and characterize. Many bioinformatic and deep learning approaches have leveraged the abundance of genomic data to recognize BGCs in bacterial genomes. However, the characterization of BGC properties remains the main bottleneck in identifying novel BGCs and their natural products. In this paper, we present a self-supervised masked language model that learns meaningful representations of BGCs with improved downstream detection and classification.

The mosquito Aedes aegypti is the vector of a number of medically-important viruses, including dengue virus, yellow virus, chikungunya virus, and Zika virus, and as such vector control is a key approach to managing the diseases they cause. Understanding the impact of vector control on these diseases is aided by first understanding its impact on Ae. aegypti population dynamics. A number of detail-rich models have been developed to couple the dynamics of the immature and adult stages of Ae. aegypti. The numerous assumptions of these models enable them to realistically characterize impacts of mosquito control, but they also constrain the ability of such models to reproduce empirical patterns that do not conform to the models’ behavior. In contrast, statistical models afford sufficient flexibility to extract nuanced signals from noisy data, yet they have limited ability to make predictions about impacts of mosquito control on disease caused by pathogens that the mosquitoes transmit without extensive data on mosquitoes and disease.Here, we demonstrate how the differing strengths of mechanistic realism and statistical flexibility can be fused into a single model. Our analysis utilizes data from 176,352 household-level Ae. aegypti aspirator collections conducted during 1999-2011 in Iquitos, Peru. The key step in our approach is to calibrate a single parameter of the model to spatio-temporal abundance patterns predicted by a generalized additive model (GAM). In effect, this calibrated parameter absorbs residual variation in the abundance time-series not captured by other features of the mechanistic model. We then used this calibrated parameter and the literature-derived parameters in the agent-based model to explore Ae. aegypti population dynamics and the impact of insecticide spraying to kill adult mosquitoes.The baseline abundance predicted by the agent-based model closely matched that predicted by the GAM. Following spraying, the agent-based model predicted that mosquito abundance rebounds within about two months, commensurate with recent experimental data from Iquitos. Our approach was able to accurately reproduce abundance patterns in Iquitos and produce a realistic response to adulticide spraying, while retaining sufficient flexibility to be applied across a range of settings.Author SummaryThe mosquito Aedes aegypti is the vector for a number of the most medically important viruses, including dengue, Zika, chikungunya, and yellow fever. Understanding the population dynamics of this mosquito, and how those dynamics might respond to vector control interventions, is critical to inform the deployment of such interventions. One of the best ways to gain this understanding is through modeling of population dynamics. Such models are often categorizes as either statistical or dynamical, and each of these approaches has advantages and disadvantages – for instance, statistical models may more closely match patterns observed in empirical data, while dynamical models are better able to predict the impact of counterfactual situations such as vector control strategies. In this paper, we present an approach which fuses these two approaches in order to gain the advantages of both: it fits empirical data on Aedes aegypti population dynamics well, while producing realistic responses to vector control interventions. Our approach has the potential to inform and improve the deployment of vector control interventions, and, when used in concert with and epidemiological model, to help reduce the burden of the diseases spread by such vectors.

Multi-view data can be generated from diverse sources, by different technologies, and in multiple modalities. In various fields, integrating information from multi-view data has pushed the frontier of discovery. In this paper, we develop a new approach for multi-view clustering, which overcomes the limitations of existing methods such as the need of pooling data across views, restrictions on the clustering algorithms allowed within each view, and the disregard for complementary information between views. Our new method, called CPS-merge analysis, merges clusters formed by the Cartesian product of single-view cluster labels, guided by the principle of maximizing clustering stability as evaluated by CPS analysis. In addition, we introduce measures to quantify the contribution of each view to the formation of any cluster. CPS-merge analysis can be easily incorporated into an existing clustering pipeline because it only requires single-view cluster labels instead of the original data. We can thus readily apply advanced single-view clustering algorithms. Importantly, our approach accounts for both consensus and complementary effects between different views, whereas existing ensemble methods focus on finding a consensus for multiple clustering results, implying that results from different views are variations of one clustering structure. Through experiments on single-cell datasets, we demonstrate that our approach frequently outperforms other state-of-the-art methods.

The long-term behaviors of biochemical systems are often described by their steady states. Deriving these states directly for complex networks arising from real-world applications, however, is often challenging. Recent work has consequently focused on network-based approaches. Specifically, biochemical reaction networks are transformed into weakly reversible and deficiency zero generalized networks, which allows the derivation of their analytic steady states. Identifying this transformation, however, can be challenging for large and complex networks. In this paper, we address this difficulty by breaking the complex network into smaller independent subnetworks and then transforming the subnetworks to derive the analytic steady states of each subnetwork. We show that stitching these solutions together leads to the analytic steady states of the original network. To facilitate this process, we develop a user-friendly and publicly available package, COMPILES (COMPutIng anaLytic stEady States). With COMPILES, we can easily test the presence of bistability of a CRISPRi toggle switch model, which was previously investigated via tremendous number of numerical simulations and within a limited range of parameters. Furthermore, COMPILES can be used to identify absolute concentration robustness (ACR), the property of a system that maintains the concentration of particular species at a steady state regardless of any initial concentrations. Specifically, our approach completely identifies all the species with and without ACR in a complex insulin model. Our method provides an effective approach to analyzing and understanding complex biochemical systems.

We introduce a new spatial statistic, the weighted pair correlation function (wPCF). The wPCF extends the existing pair correlation function (PCF) to describe spatial relationships between points marked with combinations of discrete and continuous labels. We validate its use through application to an agent-based model (ABM) which simulates interactions between macrophages and tumour cells. These interactions are influenced by the spatial positions of the cells and by macrophage phenotype, a continuous variable that ranges from anti-tumour to pro-tumour. By varying model parameters that regulate macrophage phenotype, we show that the ABM exhibits behaviours which resemble the ‘three Es of cancer immunoediting’: Equilibrium, Escape, and Elimination.We use the wPCF to analyse synthetic images generated by the ABM. We show that the wPCF generates a ‘human readable’ statistical summary of where macrophages with different phenotypes are located relative to both blood vessels and tumour cells. In combination with the cross-PCF (describing interactions between vessels and tumour cells), we show further that each of the three Es of immunoediting is characterised by a distinct ‘PCF signature’. By applying dimension reduction techniques to this signature, we identify its key features and train a support vector machine classifier to distinguish between simulation outputs based on their PCF signature. This proof-of-concept study shows how multiple spatial statistics can be combined to analyse the complex spatial features that the ABM generates, and to partition them into interpretable groups.The intricate spatial features produced by the ABM are similar to those generated by state-of-the-art multiplex imaging techniques which distinguish the spatial distribution and intensity of multiple biomarkers in biological tissue regions. Applying methods such as the wPCF to multiplex imaging data would exploit the continuous variation in biomarker intensities and generate more detailed characterisation of the spatial and phenotypic heterogeneity in tissue samples.Author summaryMultiplex images provide exquisitely detailed information about the spatial distribution and intensity of up to 40 biomarkers within two-dimensional tissue regions, creating challenges and opportunities for quantitative analysis. Although stain intensities are measured on a continuous scale, they are typically converted into discrete labels to simplify subsequent spatial analysis. In this paper we propose a new spatial statistic, the weighted pair correlation function (wPCF), which exploits, rather than neglects, the continuous variation in stain intensity contained in multiplex images, and can characterise both spatial and phenotypic heterogeneity.As proof-of-principle, we apply the wPCF to synthetic data that resemble multiplex images of solid tumours. We generate data from an agent-based model (ABM) that simulates macrophage-tumour interactions. The wPCF shows how the continuous label describing macrophage phenotype is spatially related to categorical labels associated with tumour cells and blood vessels. We demonstrate that correlation functions can categorise spatial relationships in a manner which is interpretable and quantitative.The methods we present can be used to analyse both ABM simulations and multiplex imaging data, with applications that go beyond macrophage phenotype to include other biological processes that exhibit continuous variation (e.g., cancer cell stemness, biomarkers for T-cell exhaustion, and levels of oxygenation).

Accurate prediction of synergistic effects of drug combinations can reduce the experimental costs for drug development and facilitate the discovery of novel efficacious combination therapies for clinical studies. The drug combinations with high synergy scores are regarded as synergistic ones, while those with moderate or low synergy scores are additive or antagonistic ones. The existing methods usually exploit the synergy data from the aspect of synergistic drug combinations, paying little attention to the additive or antagonistic ones. Also, they usually do not leverage the common patterns of drug combinations across different cell lines. In this paper, we propose a multi-channel graph autoencoder (MGAE)-based method for predicting the synergistic effects of drug combinations (DC), and shortly denote it as MGAE-DC. A MGAE model is built to learn the drug embeddings by considering not only synergistic combinations but also additive and antagonistic ones as three input channels. The later two channels guide the model to explicitly characterize the features of non-synergistic combinations through an encoder-decoder learning process, and thus the drug embeddings become more discriminative between synergistic and non-synergistic combinations. In addition, an attention mechanism is incorporated to fuse each cell-line’s drug embeddings across various cell lines, and a common drug embedding is extracted to capture the invariant patterns by developing a set of cell-line shared decoders. The generalization performance of our model is further improved with the invariant patterns. With the cell-line specific and common drug embeddings, our method is extended to predict the synergy scores of drug combinations by a neural network module. Experiments on four benchmark datasets demonstrate that MGAE-DC consistently outperforms the state-of-the-art methods. In-depth literature survey is conducted to find that many drug combinations predicted by MGAE-DC are supported by previous experimental studies. The source code and data are available at https://github.com/yushenshashen/MGAE-DC.

Segmenting three dimensional microscopy images is essential for understanding phenomena like morphogenesis, cell division, cellular growth and genetic expression patterns. Recently, deep learning (DL) pipelines have been developed which claim to provide high accuracy segmentation of cellular images and are increasingly considered as the state-of-the-art for image segmentation problems. However, it remains difficult to define their relative performances as the concurrent diversity and lack of uniform evaluation strategies makes it difficult to know how their results compare. In this paper, we first made an inventory of the available DL methods for 3 dimensional (3D) cell segmentation. We next implemented and quantitatively compared a number of representative DL pipelines, alongside a highly efficient non-DL method named MARS. The DL methods were trained on a common dataset of 3D cellular confocal microscopy images. Their segmentation accuracies were also tested in the presence of different image artifacts. A specific method for segmentation quality evaluation was adopted which isolates segmentation errors due to under/over segmentation. This is complemented with a 3D visualization strategy for interactive exploration of segmentation quality. Our analysis shows that the DL pipelines have different levels of accuracy. Two of them, which are end to end 3D and were originally designed for cell boundary detection, show high performance, and offer clear advantages in terms of adaptability to new data.Author summaryIn recent years a number of deep learning (DL) algorithms based on computational neural networks have been developed which claim to achieve high accuracy and automatic segmentation of 3D microscopy images. Although these algorithms have received considerable attention in the literature, it is difficult to evaluate their relative performances, while it remains unclear whether they really perform better than other, more classical segmentation methods.To clarify these issues, we performed a detailed, quantitative analysis of a number of representative DL pipelines for cell instance segmentation from 3D confocal microscopy image datasets. We developed a protocol for benchmarking the performances of such DL based segmentation pipelines using common training and test datasets, evaluation metrics and visualizations. Using this protocol, we evaluated and compared four different DL pipelines to identify their strengths and limitations. A high performance non-DL method was also included in the evaluation. We show that DL pipelines may show significant differences in their performances depending on their model architecture and pipeline components but overall show excellent adaptability to unseen data. We also show that our benchmarking protocol could be extended to a variety of segmentation pipelines and datasets.

With the recent approval by the FDA of the first disease-modifying drug for Alzheimer’s Disease (AD), personalized medicine will be increasingly important for appropriate management and counseling of patients with AD and those at risk. The growing availability of clinical biomarker data and data-driven computational modeling techniques provide an opportunity for new approaches to individualized AD therapeutic planning. In this paper, we develop a new mathematical model, based on AD cognitive, cerebrospinal fluid (CSF) and MRI biomarkers, to provide a personalized optimal treatment plan for individuals. This model is parameterized by biomarker data from the AD Neuroimaging Initiative (ADNI) cohort, a large multi-institutional database monitoring the natural history of subjects with AD and mild cognitive impairment (MCI). Optimal control theory is used to incorporate time-varying treatment controls and side-effects into the model, based on recent clinical trial data, to provide a personalized treatment regimen with anti-amyloid-beta therapy. In-silico treatment studies were conducted on the approved treatment, aducanumab, as well as on another promising anti-amyloid-beta therapy under evaluation, donanemab. Clinical trial simulations were conducted over both short-term (78 weeks) and long-term (10 years) periods with low-dose (6 mg/kg) and high-dose (10 mg/kg) regimens for aducanumab, and a single-dose regimen (1400 mg) for donanemab. Results confirm those of actual clinical trials showing a large and sustained effect of both aducanumab and donanemab on amyloid beta clearance. The effect on slowing cognitive decline was modest for both treatments, but greater for donanemab. This optimal treatment computational modeling framework can be applied to other single and combination treatments for both prediction and optimization, as well as incorporate new clinical trial data as it becomes available.

The complexity of entire T cell receptor (TCR) repertoires makes their comparison a difficult but important task. Current methods of TCR repertoire comparison can incur a high loss of distributional information by considering overly simplistic sequence- or repertoire-level characteristics. Optimal transport methods form a suitable approach for such comparison given some distance or metric between values in the sample space, with appealing theoretical and computational properties. In this paper we introduce a nonparametric approach to comparing empirical TCR repertoires that applies the Sinkhorn distance, a fast, contemporary optimal transport method, and a recently-created distance between TCRs called TCRdist. We show that our methods identify meaningful differences between samples from distinct TCR distributions for several case studies, and compete with more complicated methods despite minimal modeling assumptions and a simpler pipeline.

A key challenge in mobilising growing numbers of digitised biological specimens for scientific research is finding high-throughput methods to extract phenotypic measurements on these datasets. In this paper, we test a pose estimation approach based on Deep Learning capable of accurately placing point labels to identify key locations on specimen images. We then apply the approach to two distinct challenges that each requires identification of key features in a 2D image: (i) identifying body region-specific plumage colouration on avian specimens and (ii) measuring morphometric shape variation in Littorina snail shells. For the avian dataset, 95% of images are correctly labelled and colour measurements derived from these predicted points are highly correlated with human-based measurements. For the Littorina dataset, more than 95% of landmarks were accurately placed relative to expert-labelled landmarks and predicted landmarks reliably captured shape variation between two distinct shell ecotypes (‘crab’ vs ‘wave’). Overall, our study shows that pose estimation based on Deep Learning can generate high-quality and high-throughput point-based measurements for digitised image-based biodiversity datasets and could mark a step change in the mobilisation of such data. We also provide general guidelines for using pose estimation methods on large-scale biological datasets.

The neurovascular and neurometabolic couplings (NVC and NMC) connect cerebral activity, blood flow, and metabolism. This interconnection is used in for instance functional imaging, which analyses the blood-oxygen-dependent (BOLD) signal. The mechanisms underlying the NVC are complex, which warrants a model-based analysis of data. We have previously developed a mechanistically detailed model for the NVC, and others have proposed detailed models for cerebral metabolism. However, existing metabolic models are still not fully utilizing available magnetic resonance spectroscopy (MRS) data and are not connected to detailed models for NVC. Therefore, we herein present a new model that integrates mechanistic modelling of both MRS and BOLD data. The metabolic model covers central metabolism, using a minimal set of interactions, and can describe time-series data for glucose, lactate, aspartate, and glutamate, measured after visual stimuli. Statistical tests confirm that the model can describe both estimation data and predict independent validation data, not used for model training. The interconnected NVC model can simultaneously describe BOLD data and can be used to predict expected metabolic responses in experiments where metabolism has not been measured. This model is a step towards a useful and mechanistically detailed model for cerebral blood flow and metabolism, with potential applications in both basic research and clinical applications. Author summary The neurovascular and neurometabolic couplings are highly central for several clinical imaging techniques since these frequently use blood oxygenation (the BOLD signal) as a proxy for neuronal activity. This relationship is described by the highly complex neurovascular and neurometabolic couplings, which describe the balancing between increased metabolic demand and blood flow, and which involve several cell types and regulatory systems, which all change dynamically over time. While there are previous works that describe the neurovascular coupling in detail, neither we nor others have developed connections to corresponding mechanistic models for the third aspect, the metabolic aspect. Furthermore, magnetic resonance spectroscopy (MRS) data for such modelling readily is available. In this paper we present a minimal mechanistic model that can describe the metabolic response to visual stimuli. The model is trained to describe experimental data for the relative change in metabolic concentrations of several metabolites in the visual cortex during stimulation. The model is also validated against independent validation data, that was not used for model training. Finally, we also connect this metabolic model to a detailed mechanistic model of the neurovascular coupling. Showing that the model can describe both the metabolic response and a neurovascular response simultaneously. Funding Agencies|Swedish Research Council [2018-05418, 2018-03319, 2018-03391]; CENIIT [15.09]; Swedish foundation [ITM17-0245]; SciLifeLab and KAW [2020.0182]; H2020 project PRECISE4Q [777107]; Swedish Fund [F2019-0010]; Swedish Brain Foundation; ELLIIT; VisualSweden; VINNOVA [2020-04711]

Quantitative stochastic models of gene regulatory networks are important tools for studying cellular regulation. Such models can be formulated at many different levels of fidelity. A practical challenge is to determine what model fidelity to use in order to get accurate and representative results. The choice is important, because models of successively higher fidelity come at a rapidly increasing computational cost. In some situations, the level of detail is clearly motivated by the question under study. In many situations however, many model options could qualitatively agree with available data, depending on the amount of data and the nature of the observations. Here, an important distinction is whether we are interested in inferring the true (but unknown) physical parameters of the model or if it is sufficient to be able to capture and explain available data. The situation becomes complicated from a computational perspective because inference needs to be approximate. Most often it is based on likelihood-free Approximate Bayesian Computation (ABC) and here determining which summary statistics to use, as well as how much data is needed to reach the desired level of accuracy, are difficult tasks. Ultimately, all of these aspects—the model fidelity, the available data, and the numerical choices for inference—interplay in a complex manner. In this paper we develop a computational pipeline designed to systematically evaluate inference accuracy for a wide range of true known parameters. We then use it to explore inference settings for negative feedback gene regulation. In particular, we compare a detailed spatial stochastic model, a coarse-grained compartment-based multiscale model, and the standard well-mixed model, across several data-scenarios and for multiple numerical options for parameter inference. Practically speaking, this pipeline can be used as a preliminary step to guide modelers prior to gathering experimental data. By training Gaussian processes to approximate the distance function values, we are able to substantially reduce the computational cost of running the pipeline. De två sista författarna delar sistaförfattarskapet eSSENCE - An eScience Collaboration

The real-time analysis of infectious disease surveillance data is essential in obtaining situational awareness about the current dynamics of a major public health event such as the COVID-19 pandemic. This analysis of e.g., time-series of reported cases or fatalities is complicated by reporting delays that lead to under-reporting of the complete number of events for the most recent time points. This can lead to misconceptions by the interpreter, for instance the media or the public, as was the case with the time-series of reported fatalities during the COVID-19 pandemic in Sweden. Nowcasting methods provide real-time estimates of the complete number of events using the incomplete time-series of currently reported events and information about the reporting delays from the past. In this paper we propose a novel Bayesian nowcasting approach applied to COVID-19-related fatalities in Sweden. We incorporate additional information in the form of time-series of number of reported cases and ICU admissions as leading signals. We demonstrate with a retrospective evaluation that the inclusion of ICU admissions as a leading signal improved the nowcasting performance of case fatalities for COVID-19 in Sweden compared to existing methods.

Identifying the interrelations among cancer driver genes and the patterns in which the driver genes get mutated is critical for understanding cancer. In this paper, we study cross-sectional data from cohorts of tumors to identify the cancer-type (or subtype) specific process in which the cancer driver genes accumulate critical mutations. We model this mutation accumulation process using a tree, where each node includes a driver gene or a set of driver genes. A mutation in each node enables its children to have a chance of mutating. This model simultaneously explains the mutual exclusivity patterns observed in mutations in specific cancer genes (by its nodes) and the temporal order of events (by its edges). We introduce a computationally efficient dynamic programming procedure for calculating the likelihood of our noisy datasets and use it to build our Markov Chain Monte Carlo (MCMC) inference algorithm, ToMExO. Together with a set of engineered MCMC moves, our fast likelihood calculations enable us to work with datasets with hundreds of genes and thousands of tumors, which cannot be dealt with using available cancer progression analysis methods. We demonstrate our method’s performance on several synthetic datasets covering various scenarios for cancer progression dynamics. Then, a comparison against two state-of-the-art methods on a moderate-size biological dataset shows the merits of our algorithm in identifying significant and valid patterns. Finally, we present our analyses of several large biological datasets, including colorectal cancer, glioblastoma, and pancreatic cancer. In all the analyses, we validate the results using a set of method-independent metrics testing the causality and significance of the relations identified by ToMExO or competing methods.

Language interfaces with many other cognitive domains. This paper explores how interactions at these interfaces can be studied with deep learning methods, focusing on the relation between language emergence and visual perception. To model the emergence of language, a sender and a receiver agent are trained on a reference game. The agents are implemented as deep neural networks, with dedicated vision and language modules. Motivated by the mutual influence between language and perception in cognition, we apply systematic manipulations to the agents’ (i) visual representations, to analyze the effects on emergent communication, and (ii) communication protocols, to analyze the effects on visual representations. Our analyses show that perceptual biases shape semantic categorization and communicative content. Conversely, if the communication protocol partitions object space along certain attributes, agents learn to represent visual information about these attributes more accurately, and the representations of communication partners align. Finally, an evolutionary analysis suggests that visual representations may be shaped in part to facilitate the communication of environmentally relevant distinctions. Aside from accounting for co-adaptation effects between language and perception, our results point out ways to modulate and improve visual representation learning and emergent communication in artificial agents.

Biological systems are noisy by nature. This aspect is reflected in our experimental measurements and should be reflected in the models we build to better understand these systems. Noise can be especially consequential when trying to interpret specific regulatory interactions, i.e. regulatory network edges. In this paper, we propose a method to explicitly encode edge-noise in Boolean dynamical systems by probabilistic edge-weight (PEW) operators. PEW operators have two important features: first, they introduce a form of edge-weight into Boolean models through the noise, second, the noise is dependent on the dynamical state of the system, which enables more biologically meaningful modeling choices. Moreover, we offer a simple-to-use implementation in the already well-established BooleanNet framework. In two application cases, we show how the introduction of just a few PEW operators in Boolean models can fine-tune the emergent dynamics and increase the accuracy of qualitative predictions. This includes fine-tuning interactions which cause non-biological behaviors when switching between asynchronous and synchronous update schemes in dynamical simulations. Moreover, PEW operators also open the way to encode more exotic cellular dynamics, such as cellular learning, and to implementing edge-weights for regulatory networks inferred from omics data.

Activities of transcription factors (TFs) are temporally modulated to regulate dynamic cellular processes, including development, homeostasis, and disease. Recent developments of bioinformatic tools have enabled the analysis of TF activities using transcriptome data. However, because these methods typically use exon-based target expression levels, the estimated TF activities have limited temporal accuracy. To address this, we proposed a TF activity measure based on intron-level information in time-series RNA-seq data, and implemented it to decode the temporal control of TF activities during dynamic processes. We showed that TF activities inferred from intronic reads can better recapitulate instantaneous TF activities compared to the exon-based measure. By analyzing public and our own time-series transcriptome data, we found that intron-based TF activities improve the characterization of temporal phasing of cycling TFs during circadian rhythm, and facilitate the discovery of two temporally opposing TF modules during T cell activation. Collectively, we anticipate that the proposed approach would be broadly applicable for decoding global transcriptional architecture during dynamic processes., Author summary Many health-related cellular processes, such as immune response and disease progression, involve dynamic changes of gene expression state, which are orchestrated by transcription factors. Dissecting the activities of transcription factors is thus important for understanding cellular processes and for interfering with dysregulated processes. Our ability to analyze transcription factor activities has been facilitated by genome-wide gene expression data from high-throughput assays such as RNA sequencing. Existing methods typically estimate transcription factor activities based on the expression levels of matured mRNAs of target genes. However, because the levels of matured mRNAs are affected by transcriptional and post-transcriptional regulatory activities, the estimated transcription factor activities may not faithfully recapitulate the regulatory activities of transcription factors. In this paper, we proposed and validated an alternative approach for analyzing transcription factor activities using the expression levels of unmatured mRNAs of target genes, allowing us to decode how transcription factor activities are temporally controlled during key biological processes. Our results provide insights into the temporal phasing of key circadian regulator activities in mouse liver, and uncover two temporally opposing modules of transcription factors that dictate the immune responses in T cells. Therefore, this approach can help understand the regulatory principles of dynamic cellular processes.

The aim of this paper is two-fold. First, we propose a new computational method to investigate the particularities of evolution. Second, we apply this method to a model of gene regulatory networks (GRNs) and explore the evolution of mutational robustness and bistability. Living systems have developed their functions through evolutionary processes. To understand the particularities of this process theoretically, evolutionary simulation (ES) alone is insufficient because the outcomes of ES depend on evolutionary pathways. We need a reference system for comparison. An appropriate reference system for this purpose is an ensemble of the randomly sampled genotypes. However, generating high-fitness genotypes by simple random sampling is difficult because such genotypes are rare. In this study, we used the multicanonical Monte Carlo method developed in statistical physics to construct a reference ensemble of GRNs and compared it with the outcomes of ES. We obtained the following results. First, mutational robustness was significantly higher in ES than in the reference ensemble at the same fitness level. Second, the emergence of a new phenotype, bistability, was delayed in evolution. Third, the bistable group of GRNs contains many mutationally fragile GRNs compared with those in the non-bistable group. This suggests that the delayed emergence of bistability is a consequence of the mutation-selection mechanism., Comment: 14 pages, 12 figures

Workflows have become a core part of computational scientific analysis in recent years. Automated computational workflows multiply the power of researchers, potentially turning “hand-cranked” data processing by informaticians into robust factories for complex research output.However, in order for a piece of software to be usable as a workflow-ready tool, it may require alteration from its likely origin as a standalone tool. Research software is often created in response to the need to answer a research question with the minimum expenditure of time and money in resource-constrained projects. The level of quality might range from “it works on my computer” to mature and robust projects with support across multiple operating systems.Despite significant increase in uptake of workflow tools, there is little specific guidance for writing software intended to slot in as a tool within a workflow; or on converting an existing standalone research-quality software tool into a reusable, composable, well-behaved citizen within a larger workflow.In this paper we present 10 simple rules for how a software tool can be prepared for workflow use.

Constructing biological networks capable of performing specific biological functionalities has been of sustained interest in synthetic biology. Adaptation is one such ubiquitous functional property, which enables every living organism to sense a change in its surroundings and return to its operating condition prior to the disturbance. In this paper, we present a generic systems theory-driven method for designing adaptive protein networks. First, we translate the necessary qualitative conditions for adaptation to mathematical constraints using the language of systems theory, which we then map back as ‘design requirements’ for the underlying networks. We go on to prove that a protein network with different input–output nodes (proteins) needs to be at least of third-order in order to provide adaptation. Next, we show that the necessary design principles obtained for a three-node network in adaptation consist of negative feedback or a feed-forward realization. We argue that presence of a particular class of negative feedback or feed-forward realization is necessary for a network of any size to provide adaptation. Further, we claim that the necessary structural conditions derived in this work are the strictest among the ones hitherto existed in the literature. Finally, we prove that the capability of producing adaptation is retained for the admissible motifs even when the output node is connected with a downstream system in a feedback fashion. This result explains how complex biological networks achieve robustness while keeping the core motifs unchanged in the context of a particular functionality. We corroborate our theoretical results with detailed and thorough numerical simulations. Overall, our results present a generic, systematic and robust framework for designing various kinds of biological networks., Author summary Biological systems display a remarkable diversity of functionalities, many of which can be conceived as the response of a large network composed of small interconnecting modules. Unravelling the connection pattern, i.e. design principles, behind important biological functionalities is one of the most challenging problems in systems biology. One such phenomenon is perfect adaptation, which merits special attention owing to its universal presence ranging from chemotaxis in bacterial cells to calcium homeostasis in mammalian cells. The present work focuses on finding the design principles for perfect adaptation in the presence of a stair-case type disturbance. To this end, the current work proposes a systems-theoretic approach to deduce precise mathematical (hence structural) conditions that comply with the key performance parameters for adaptation. The approach is agnostic to the particularities of the reaction kinetics, underlining the dominant role of the topological structure on the response of the network. Notably, the design principles obtained in this work serve as the most strict necessary structural conditions for a network of any size to provide perfect adaptation.

Experiments on tumor spheroids have shown that compressive stress from their environment can reversibly decrease tumor expansion rates and final sizes. Stress release experiments show that nonuniform anisotropic elastic stresses can be distributed throughout. The elastic stresses are maintained by structural proteins and adhesive molecules, and can be actively relaxed by a variety of biophysical processes. In this paper, we present a new continuum model to investigate how the growth-induced elastic stresses and active stress relaxation, in conjunction with cell size control feedback machinery, regulate the cell density and stress distributions within growing tumors as well as the tumor sizes in the presence of external physical confinement and gradients of growth-promoting chemical fields. We introduce an adaptive reference map that relates the current position with the reference position but adapts to the current position in the Eulerian frame (lab coordinates) via relaxation. This type of stress relaxation is similar to but simpler than the classical Maxwell model of viscoelasticity in its formulation. By fitting the model to experimental data from two independent studies of tumor spheroid growth and their cell density distributions, treating the tumors as incompressible, neo-Hookean elastic materials, we find that the rates of stress relaxation of tumor tissues can be comparable to volumetric growth rates. Our study provides insight on how the biophysical properties of the tumor and host microenvironment, mechanical feedback control and diffusion-limited differential growth act in concert to regulate spatial patterns of stress and growth. When the tumor is stiffer than the host, our model predicts tumors are more able to change their size and mechanical state autonomously, which may help to explain why increased tumor stiffness is an established hallmark of malignant tumors., Author summary The mechanical state of cells can modulate their growth and division dynamics via mechanotransduction, which affects both the cell size distribution and the tissue size as a whole. Experiments on tumor spheroids have shown that compressive stress from their environment can reversibly decrease tumor expansion rates and final sizes. Besides external confinement and compression on the tumor border, a heterogeneous stress field can be generated inside the tumor by nutrient-driven differential growth. Such growth-induced mechanical stresses can be relaxed by tissue rearrangement, which happens during cell neighbor exchanges, cell divisions, and extracellular matrix renewal. In this study, we have developed a continuum model that describes the above mechanical interactions and the dynamics of tissue rearrangement explicitly. Motivated by published experimental data, we consider mechanotransduction where the local compressive stress slows down cell growth and cell size reduction limits cell division. We have analyzed how external mechanical stimuli and internal processes influence the outcome of cell-and-tissue sizes and spatial patterns of cell density and mechanical stress in growing tumors.

Background Sulfadoxine-pyrimethamine (SP) is recommended in Africa in several antimalarial preventive regimens including Intermittent Preventive Treatment in pregnant women (IPTp), Intermittent Preventive Treatment in infants (IPTi) and Seasonal Malaria Chemoprevention (SMC). The effectiveness of SP-based preventive treatments are threatened in areas where Plasmodium falciparum resistance to SP is high. The prevalence of mutations in the dihydropteroate synthase gene (pfdhps) can be used to monitor SP effectiveness. IPTi-SP is recommended only in areas where the prevalence of the pfdhps540E mutation is below 50%. It has also been suggested that IPTp-SP does not have a protective effect in areas where the pfdhps581G mutation, exceeds 10%. However, pfdhps mutation prevalence data in Africa are extremely heterogenous and scattered, with data completely missing from many areas. Methods and findings The WWARN SP Molecular Surveyor database was designed to summarize dihydrofolate reductase (pfdhfr) and pfdhps gene mutation prevalence data. In this paper, pfdhps mutation prevalence data was used to generate continuous spatiotemporal surface maps of the estimated prevalence of the SP resistance markers pfdhps437G, pfdhps540E, and pfdhps581G in Africa from 1990 to 2020 using a geostatistical model, with a Bayesian inference framework to estimate uncertainty. The maps of estimated prevalence show an expansion of the pfdhps437G mutations across the entire continent over the last three decades. The pfdhps540E mutation emerged from limited foci in East Africa to currently exceeding 50% estimated prevalence in most of East and South East Africa. pfdhps540E distribution is expanding at low or moderate prevalence in central Africa and a predicted focus in West Africa. Although the pfdhps581G mutation spread from one focus in East Africa in 2000, to exceeding 10% estimated prevalence in several foci in 2010, the predicted distribution of the marker did not expand in 2020, however our analysis indicated high uncertainty in areas where pfdhps581G is present. Uncertainty was higher in spatial regions where the prevalence of a marker is intermediate or where prevalence is changing over time. Conclusions The WWARN SP Molecular Surveyor database and a set of continuous spatiotemporal surface maps were built to provide users with standardized, current information on resistance marker distribution and prevalence estimates. According to the maps, the high prevalence of pfdhps540E mutation was to date restricted to East and South East Africa, which is reassuring for continued use of IPTi and SMC in West Africa, but continuous monitoring is needed as the pfdhps540E distribution is expanding. Several foci where pfdhps581G prevalence exceeded 10% were identified. More data on the pfdhps581G distribution in these areas needs to be collected to guide IPTp-SP recommendations. Prevalence and uncertainty maps can be utilized together to strategically identify sites where increased surveillance can be most informative. This study combines a molecular marker database and predictive modelling to highlight areas of concern, which can be used to support decisions in public health, highlight knowledge gaps in certain regions, and guide future research.

In this paper, a logical-based mathematical model of the cellular pathways involved in the COVID-19 infection has been developed to study various drug treatments (single or in combination), in different illness scenarios, providing insights into their mechanisms of action. Drug simulations suggest that the effects of single drugs are limited, or depending on the scenario counterproductive, whereas better results appear combining different treatments. Specifically, the combination of the anti-inflammatory Baricitinib and the anti-viral Remdesivir showed significant benefits while a stronger efficacy emerged from the triple combination of Baricitinib, Remdesivir, and the corticosteroid Dexamethasone. Together with a sensitivity analysis, we performed an analysis of the mechanisms of the drugs to reveal their impact on molecular pathways.

Stay-at-home orders and shutdowns of non-essential businesses are powerful, but socially costly, tools to control the pandemic spread of SARS-CoV-2. Mass testing strategies, which rely on widely administered frequent and rapid diagnostics to identify and isolate infected individuals, could be a potentially less disruptive management strategy, particularly where vaccine access is limited. In this paper, we assess the extent to which mass testing and isolation strategies can reduce reliance on socially costly non-pharmaceutical interventions, such as distancing and shutdowns. We develop a multi-compartmental model of SARS-CoV-2 transmission incorporating both preventative non-pharmaceutical interventions (NPIs) and testing and isolation to evaluate their combined effect on public health outcomes. Our model is designed to be a policy-guiding tool that captures important realities of the testing system, including constraints on test administration and non-random testing allocation. We show how strategic changes in the characteristics of the testing system, including test administration, test delays, and test sensitivity, can reduce reliance on preventative NPIs without compromising public health outcomes in the future. The lowest NPI levels are possible only when many tests are administered and test delays are short, given limited immunity in the population. Reducing reliance on NPIs is highly dependent on the ability of a testing program to identify and isolate unreported, asymptomatic infections. Changes in NPIs, including the intensity of lockdowns and stay at home orders, should be coordinated with increases in testing to ensure epidemic control; otherwise small additional lifting of these NPIs can lead to dramatic increases in infections, hospitalizations and deaths. Importantly, our results can be used to guide ramp-up of testing capacity in outbreak settings, allow for the flexible design of combined interventions based on social context, and inform future cost-benefit analyses to identify efficient pandemic management strategies., Author summary The global spread of SARS-CoV-2 and the strategies used to manage it have come at significant societal costs. We analyze how mixed control strategies, which utilize interventions that prevent new infections from occurring (e.g., distancing or shut-downs) and others that actively search for and isolate existing infections (here, mass testing), can achieve improved public health outcomes while avoiding severe socio-economic burdens. Our results suggest that increasing testing capacity, including the number of tests available and the speed at which test results are provided, can reduce reliance on costly preventative interventions. Such reduction is possible with more isolation of active infections, including those without reported symptoms. However, failing to maintain preventative interventions without sufficient testing capacity can lead to large increases in infection burdens. By defining the combined effect of these interventions through mathematical models, this study provides insight into relaxation of distancing measures, and lays the groundwork for future public health economics analyses on the cost-effectiveness of combined management strategies.

The spread of infectious diseases such as COVID-19 presents many challenges to healthcare systems and infrastructures across the world, exacerbating inequalities and leaving the world’s most vulnerable populations most affected. Given their density and available infrastructure, refugee and internally displaced person (IDP) settlements can be particularly susceptible to disease spread. In this paper we present an agent-based modeling approach to simulating the spread of disease in refugee and IDP settlements under various non-pharmaceutical intervention strategies. The model, based on the June open-source framework, is informed by data on geography, demographics, comorbidities, physical infrastructure and other parameters obtained from real-world observations and previous literature. The development and testing of this approach focuses on the Cox’s Bazar refugee settlement in Bangladesh, although our model is designed to be generalizable to other informal settings. Our findings suggest the encouraging self-isolation at home of mild to severe symptomatic patients, as opposed to the isolation of all positive cases in purpose-built isolation and treatment centers, does not increase the risk of secondary infection meaning the centers can be used to provide hospital support to the most intense cases of COVID-19. Secondly we find that mask wearing in all indoor communal areas can be effective at dampening viral spread, even with low mask efficacy and compliance rates. Finally, we model the effects of reopening learning centers in the settlement under various mitigation strategies. For example, a combination of mask wearing in the classroom, halving attendance regularity to enable physical distancing, and better ventilation can almost completely mitigate the increased risk of infection which keeping the learning centers open may cause. These modeling efforts are being incorporated into decision making processes to inform future planning, and further exercises should be carried out in similar geographies to help protect those most vulnerable., Author summary The spread of infectious diseases presents many challenges to healthcare systems and infrastructures across the world. Given their density and available infrastructure, refugee and internally displaced person (IDP) settlements can be particularly susceptible to the dangers of disease spread. This study seeks to understand how COVID-19 spreads in settlements, focusing on the Cox’s Bazar refugee settlement in Bangladesh. Our model simulates the movements and interactions of each individual in the settlement, incorporating information about family structures and demographic attributes, to understand how COVID-19 might spread under various intervention strategies. Our analysis suggests that mask wearing in indoor locations can have a significant effect on disease spread, even when wearing reusable cotton masks, which the people in the settlement can make themselves. We also look at different ways to treat individuals who only have milder symptoms and don’t yet require hospitalisation, as well as various scenarios which might allow for the safe reopening of schools in the settlement. With almost 80 million forcibly displaced people in the world, we hope that this work will inspire more modeling groups to focus on these vulnerable populations, which have been traditionally under-served by such efforts, to ensure no one is left behind.

Although chemotherapy is a standard treatment for cancer, it comes with significant side effects. In particular, certain agents can induce severe muscle loss, known as cachexia, worsening patient quality of life and treatment outcomes. 5-fluorouracil, an anti-cancer agent used to treat several cancers, has been shown to cause muscle loss. Experimental data indicates a non-linear dose-dependence for muscle loss in mice treated with daily or week-day schedules. We present a mathematical model of chemotherapy-induced muscle wasting that captures this non-linear dose-dependence. Area-under-the-curve metrics are proposed to quantify the treatment’s effects on lean mass and tumour control. Model simulations are used to explore alternate dosing schedules, aging effects, and morphine use in chemotherapy treatment with the aim of better protecting lean mass while actively targeting the tumour, ultimately leading to improved personalization of treatment planning and improved patient quality of life.Author SummaryIn this paper we present a novel mathematical model for muscle loss due to cancer chemotherapy treatment. Loss of muscle mass relates to increased drug toxicity and side-effects, and to decreased patient quality of life and survival rates. With our model, we examine the therapeutic efficacy of various dosing schedules with the aim of controlling a growing tumour while also preserving lean mass. Preservation of body composition, in addition to consideration of inflammation and immune interactions, the gut microbiome, and other systemic health measures, may lead to improved patient-specific treatment plans that improve patient quality of life.

System identification techniques -- projection pursuit regression models (PPRs) and convolutional neural networks (CNNs) -- provide state-of-the-art performance in predicting visual cortical neurons' responses to arbitrary input stimuli. However, the constituent kernels recovered by these methods are often noisy and lack coherent structure, making it difficult to understand the underlying component features of a neuron's receptive field. In this paper, we show that using a dictionary of diverse kernels with complex shapes learned from natural scenes based on efficient coding theory, as the front-end for PPRs and CNNs can improve their performance in neuronal response prediction as well as algorithmic data efficiency and convergence speed. Extensive experimental results also indicate that these sparse-code kernels provide important information on the component features of a neuron's receptive field. In addition, we find that models with the complex-shaped sparse code front-end are significantly better than models with a standard orientation-selective Gabor filter front-end for modeling V1 neurons that have been found to exhibit complex pattern selectivity. We show that the relative performance difference due to these two front-ends can be used to produce a sensitive metric for detecting complex selectivity in V1 neurons., 22 pages

The modelling of many real-world problems relies on computationally heavy simulations of randomly interacting individuals or agents. However, the values of the parameters that underlie the interactions between agents are typically poorly known, and hence they need to be inferred from macroscopic observations of the system. Since statistical inference rests on repeated simulations to sample the parameter space, the high computational expense of these simulations can become a stumbling block. In this paper, we compare two ways to mitigate this issue in a Bayesian setting through the use of machine learning methods: One approach is to construct lightweight surrogate models to substitute the simulations used in inference. Alternatively, one might altogether circumvent the need for Bayesian sampling schemes and directly estimate the posterior distribution. We focus on stochastic simulations that track autonomous agents and present two case studies: tumour growths and the spread of infectious diseases. We demonstrate that good accuracy in inference can be achieved with a relatively small number of simulations, making our machine learning approaches orders of magnitude faster than classical simulation-based methods that rely on sampling the parameter space. However, we find that while some methods generally produce more robust results than others, no algorithm offers a one-size-fits-all solution when attempting to infer model parameters from observations. Instead, one must choose the inference technique with the specific real-world application in mind. The stochastic nature of the considered real-world phenomena poses an additional challenge that can become insurmountable for some approaches. Overall, we find machine learning approaches that create direct inference machines to be promising for real-world applications. We present our findings as general guidelines for modelling practitioners.

Working memory is a cognitive function involving the storage and manipulation of latent information over brief intervals of time, thus making it crucial for context-dependent computation. Here, we use a top-down modeling approach to examine network-level mechanisms of working memory, an enigmatic issue and central topic of study in neuroscience. We optimize thousands of recurrent rate-based neural networks on a working memory task and then perform dynamical systems analysis on the ensuing optimized networks, wherein we find that four distinct dynamical mechanisms can emerge. In particular, we show the prevalence of a mechanism in which memories are encoded along slow stable manifolds in the network state space, leading to a phasic neuronal activation profile during memory periods. In contrast to mechanisms in which memories are directly encoded at stable attractors, these networks naturally forget stimuli over time. Despite this seeming functional disadvantage, they are more efficient in terms of how they leverage their attractor landscape and paradoxically, are considerably more robust to noise. Our results provide new hypotheses regarding how working memory function may be encoded within the dynamics of neural circuits., Author summary The ability to remember information for brief periods of time before using it is a key human ability. For example, retaining a phone number for a few moments prior to entering it into a keypad. Such ability, known as working memory, enables many more complex functions such as planning and reasoning. In this paper, we use theory and computational modeling approaches to try and better understand how circuits and networks in the brain might be achieving working memory. Specifically, we construct hypothetical network models to perform a task that embodies essential aspects of working memory. We then dissect our model to reveal how its components—simulated neural units—interact with each other to represent and maintain information. It turns out that some models can achieve memory by maintaining a fixed representation over time, i.e., the units remain ‘still’ during memory. However, other models fluctuate their activity during memory in a particular way that is seemingly quite efficient and less susceptible to distraction. In total, our computational study provides new theory for how this form of memory might be implemented in brain networks.

The metagenome embedded in urban sewage is an attractive new data source to understand urban ecology and assess human health status at scales beyond a single host. Analyzing the viral fraction of wastewater in the ongoing COVID-19 pandemic has shown the potential of wastewater as aggregated samples for early detection, prevalence monitoring, and variant identification of human diseases in large populations. However, using census-based population size instead of real-time population estimates can mislead the interpretation of data acquired from sewage, hindering assessment of representativeness, inference of prevalence, or comparisons of taxa across sites. Here, we show that taxon abundance and sub-species diversisty in gut-associated microbiomes are new feature space to utilize for human population estimation. Using a population-scale human gut microbiome sample of over 1,100 people, we found that taxon-abundance distributions of gut-associated multi-person microbiomes exhibited generalizable relationships with respect to human population size. Here and throughout this paper, the human population size is essentially the sample size from the wastewater sample. We present a new algorithm, MicrobiomeCensus, for estimating human population size from sewage samples. MicrobiomeCensus harnesses the inter-individual variability in human gut microbiomes and performs maximum likelihood estimation based on simultaneous deviation of multiple taxa’s relative abundances from their population means. MicrobiomeCensus outperformed generic algorithms in data-driven simulation benchmarks and detected population size differences in field data. New theorems are provided to justify our approach. This research provides a mathematical framework for inferring population sizes in real time from sewage samples, paving the way for more accurate ecological and public health studies utilizing the sewage metagenome.

Protein structure alignment algorithms are often time-consuming, resulting in challenges for large-scale protein structure similarity-based retrieval. There is an urgent need for more efficient structure comparison approaches as the number of protein structures increases rapidly. In this paper, we propose an effective graph-based protein structure representation learning method, GraSR, for fast and accurate structure comparison. In GraSR, a graph is constructed based on the intra-residue distance derived from the tertiary structure. Then, deep graph neural networks (GNNs) with a short-cut connection learn graph representations of the tertiary structures under a contrastive learning framework. To further improve GraSR, a novel dynamic training data partition strategy and length-scaling cosine distance are introduced. We objectively evaluate our method GraSR on SCOPe v2.07 and a new released independent test set from PDB database with a designed comprehensive performance metric. Compared with other state-of-the-art methods, GraSR achieves about 7%-10% improvement on two benchmark datasets. GraSR is also much faster than alignment-based methods. We dig into the model and observe that the superiority of GraSR is mainly brought by the learned discriminative residue-level and global descriptors. The web-server and source code of GraSR are freely available at www.csbio.sjtu.edu.cn/bioinf/GraSR/ for academic use.

Animals rely on sensory feedback to generate accurate, reliable movements. In many flying insects, strain-sensitive neurons on the wings provide rapid feedback that is critical for stable flight control. While the impacts of wing structure on aerodynamic performance have been widely studied, the impacts of wing structure on sensing are largely unexplored. In this paper, we show how the structural properties of the wing and encoding by mechanosensory neurons interact to jointly determine optimal sensing strategies and performance. Specifically, we examine how neural sensors can be placed effectively on a flapping wing to detect body rotation about different axes, using a computational wing model with varying flexural stiffness. A small set of mechanosensors, conveying strain information at key locations with a single action potential per wingbeat, enable accurate detection of body rotation. Optimal sensor locations are concentrated at either the wing base or the wing tip, and they transition sharply as a function of both wing stiffness and neural threshold. Moreover, the sensing strategy and performance is robust to both external disturbances and sensor loss. Typically, only five sensors are needed to achieve near-peak accuracy, with a single sensor often providing accuracy well above chance. Our results show that small-amplitude, dynamic signals can be extracted efficiently with spatially and temporally sparse sensors in the context of flight. The demonstrated interaction of wing structure and neural encoding properties points to the importance of understanding each in the context of their joint evolution., Author summary In addition to generating forces for flight, insect wings also serve an important role as sensory structures, providing rapid feedback about wing bending that is used to stabilize flight. While much is known about how wing structure affects aerodynamic performance, the effects of wing structure on sensing remain unexplored. Using a computational model of a flapping wing, we examine how sensing strategies depend on wing stiffness and sensor properties. We show that body rotations can be accurately detected with a small number of sensors on the wing across a wide range of conditions. Optimal sensor locations are clustered at either the wing base or wing tip, depending on a combination of wing stiffness and sensor properties. Moreover, sensing performance is robust to multiple kinds of perturbations. Our work provides a basis for understanding how wing structure impacts incoming sensory information during flight.

© 2021 Moore et al., Bioinformatics training is required at every stage of a scientist’s research career. Continual bioinformatics training allows exposure to an ever-changing and growing repertoire of techniques and databases, and so biologists, computational scientists, and healthcare practitioners are all seeking learning opportunities in the use of computational resources and tools designed for data storage, retrieval, and analysis. There are abundant opportunities for accessing bioinformatics training for scientists in high-income countries (HICs), with well-equipped facilities and participants and trainers requiring minimal travel and financial costs alongside a range of general advice for developing short bioinformatics training courses [1–3]. However, regionally targeted bioinformatics training in low- and middle-income countries (LMICs) often requires more extensive local and external support, organization, and travel. Due to the limited expertise in bioinformatics in LMICs in general, most bioinformatics training requires a fair amount of collaboration with experts beyond the local community, country, or region. A common model of training, used as the basis of this article, includes a local host collaborating with local, regional, and international experts gathering to train local or regional participants. Recently, there has been a growth of capacity strengthening initiatives in LMICs, such as the Pan African Bioinformatics Network for Human Heredity and Health in Africa (H3ABioNet) Initiative [4–6], the Capacity Building for Bioinformatics in Latin America (CABANA) Project [7], the Asia Pacific BioInformatics Network (APBioNet) [8], and the Wellcome Connecting Science Courses and Conferences program [9]. One of the important strands of these initiatives is a drive to organize and deliver valuable bioinformatics training, but organizing and delivering short bioinformatics training workshops in an LMIC present a unique set of challenges. This paper attempts to build upon the sage advice for organizing bioinformatics workshops with specific guidance for organizing and delivering them in LMICs. It describes the processes to follow in organizing courses taking into consideration the low-resource setting. We should also note that LMICs are not a monolithic group and that setting, context, temporality, and specific location matters. LMICs are a complex regional grouping [10] and should be treated as such; however, we will present some common lessons that we hope will help organizers and trainers of bioinformatics training events in LMICs to navigate the often different, challenging, and rewarding experience., The authors who contributed to this manuscript are funded as follows: BM receives salary support from Wellcome Trust grants [WT108749/Z/15/Z, WT108749/Z/15/A], PC, VR, NM, AG’s salaries are funded in whole, or in part, by the NIH Common Fund H3ABioNet grant [U24HG006941], MC, SLFV, AR, PG, PCL’s salaries were partly funded by the UKRI-BBSRC ‘Capacity building for bioinformatics in Latin America’ (CABANA) grant, on behalf of the Global Challenges Research Fund [BB/P027849/1], JDLR is funded by ISCiii AES [ref. PI18/00591] at the CSIC/USAL (Spain) and by CYTED, RIABIO (Red Iberoamericana 521RT0118), AM’s salary is funded by [WT206194/Z/17/Z], GO is funded by the CABANA grant and SM is funded by the EMBL-EBI.

Decision-making about pandemic mitigation often relies upon simulation modelling. Models of disease transmission through networks of contacts--between individuals or between population centres--are increasingly used for these purposes. Real-world contact networks are rich in structural features that influence infection transmission, such as tightly-knit local communities that are weakly connected to one another. In this paper, we propose a new flow-based edge-betweenness centrality method for detecting bottleneck edges that connect nodes in contact networks. In particular, we utilize convex optimization formulations based on the idea of diffusion with p-norm network flow. Using simulation models of COVID-19 transmission through real network data at both individual and county levels, we demonstrate that targeting bottleneck edges identified by the proposed method reduces the number of infected cases by up to 10% more than state-of-the-art edge-betweenness methods. Furthermore, the proposed method is orders of magnitude faster than existing methods., Comment: 38 pages, 21 figures

Recent advancements in in situ methods, such as multiplexed in situ RNA hybridization and in situ RNA sequencing, have deepened our understanding of the way biological processes are spatially organized in tissues. Automated image processing and spot-calling algorithms for analyzing in situ transcriptomics images have many parameters which need to be tuned for optimal detection. Having ground truth datasets (images where there is very high confidence on the accuracy of the detected spots) is essential for evaluating these algorithms and tuning their parameters. We present a first-in-kind open-source toolkit and framework for in situ transcriptomics image analysis that incorporates crowdsourced annotations, alongside expert annotations, as a source of ground truth for the analysis of in situ transcriptomics images. The kit includes tools for preparing images for crowdsourcing annotation to optimize crowdsourced workers’ ability to annotate these images reliably, performing quality control (QC) on worker annotations, extracting candidate parameters for spot-calling algorithms from sample images, tuning parameters for spot-calling algorithms, and evaluating spot-calling algorithms and worker performance. These tools are wrapped in a modular pipeline with a flexible structure that allows users to take advantage of crowdsourced annotations from any source of their choice. We tested the pipeline using real and synthetic in situ transcriptomics images and annotations from the Amazon Mechanical Turk system obtained via Quanti.us. Using real images from in situ experiments and simulated images produced by one of the tools in the kit, we studied worker sensitivity to spot characteristics and established rules for annotation QC. We explored and demonstrated the use of ground truth generated in this way for validating spot-calling algorithms and tuning their parameters, and confirmed that consensus crowdsourced annotations are a viable substitute for expert-generated ground truth for these purposes., Author summary To understand important biological processes such as development, wound healing, and disease, it is necessary to study where different genes are expressed in a tissue. RNA molecules can be visualized within tissues by using in situ transcriptomics tools, which use fluorescent probes that bind to specific RNA target molecules and appear in microscopy images as bright spots. Algorithms can be used to find the locations of these spots, but ground truth datasets (images with spots located to high accuracy) are needed to evaluate these algorithms and tune their parameters. While the typical way of generating ground truth datasets is having an expert annotate the spots by hand, many in situ transcriptomics image datasets are too large for this. However, it is often easy for non-experts to identify the spots with minimal training. In this paper, we present an open-source toolkit and framework that incorporates crowdsourced annotations alongside expert annotations as a source of ground truth for the analysis of in situ transcriptomics images. We explored and demonstrated the use of this framework for validating spot-calling algorithms and tuning their parameters, and confirmed that consensus crowdsourced annotations are a viable substitute for expert-generated ground truth.

Ethanol is one of the most widely used recreational substances in the world and due to its ubiquitous use, ethanol abuse has been the cause of over 3.3 million deaths each year. In addition to its effects, ethanol’s primary metabolite, acetaldehyde, is a carcinogen that can cause symptoms of facial flushing, headaches, and nausea. How strongly ethanol or acetaldehyde affects an individual depends highly on the genetic polymorphisms of certain genes. In particular, the genetic polymorphisms of mitochondrial aldehyde dehydrogenase, ALDH2, play a large role in the metabolism of acetaldehyde. Thus, it is important to characterize how genetic variations can lead to different exposures and responses to ethanol and acetaldehyde. While the pharmacokinetics of ethanol metabolism through alcohol dehydrogenase have been thoroughly explored in previous studies, in this paper, we combined a base physiologically-based pharmacokinetic (PBPK) model with a whole-body genome-scale model (WBM) to gain further insight into the effect of other less explored processes and genetic variations on ethanol metabolism. This combined model was fit to clinical data and used to show the effect of alcohol concentrations, organ damage, ALDH2 enzyme polymorphisms, and ALDH2-inhibiting drug disulfiram on ethanol and acetaldehyde exposure. Through estimating the reaction rates of auxiliary processes with dynamic Flux Balance Analysis, The PBPK-WBM was able to navigate around a lack of kinetic constants traditionally associated with PK modelling and demonstrate the compensatory effects of the body in response to decreased liver enzyme expression. Additionally, the model demonstrated that acetaldehyde exposure increased with higher dosages of disulfiram and decreased ALDH2 efficiency, and that moderate consumption rates of ethanol could lead to unexpected accumulations in acetaldehyde. This modelling framework combines the comprehensive steady-state analyses from genome-scale models with the dynamics of traditional PK models to create a highly personalized form of PBPK modelling that can push the boundaries of precision medicine., Author summary Alcohol is a widely used recreational drug in many parts of the world and it is often abused or misused, leading to the deaths of millions of people each year from driving under the influence and overdose. Additionally, the body breaks down alcohol into acetaldehyde, a carcinogen that has its own effects ranging from headaches and nausea to liver damage. The effects of ethanol and acetaldehyde vary due to genetic variations that create different forms of the enzymes responsible for breaking them down. Due to these differences, it is important to characterize how these changes affect the metabolism of alcohol and acetaldehyde. To capture these differences, we have created a new model that integrates the traditional pharmacokinetic model with a whole-body genome-scale model that can characterize different genetic variations. In addition, traditional models often require experimentally measured data, yet with this new framework we avoid this tedious process by mathematically solving the genome-scale model with the dynamic Flux Balance Analysis technique, allowing for gap filling. Through this model, we show that the whole-body genome-scale model demonstrates flexibility and robustness that has not been seen before in pharmacokinetic models. Our model combines advantages from both pharmacokinetic and genome-scale modelling and can be personalized to characterize individual reactions to other drugs and further precision medicine.