Search

Both machine learning and mechanistic modelling approaches have been used independently with great success in systems biology. Machine learning excels in deriving statistical relationships and quantitative prediction from data, while mechanistic modelling is a powerful approach to capture knowledge and infer causal mechanisms underpinning biological phenomena. Importantly, the strengths of one are the weaknesses of the other, which suggests that substantial gains can be made by combining machine learning with mechanistic modelling, a field referred to as Scientific Machine Learning (SciML). In this review we discuss recent advances in combining these two approaches for systems biology, and point out future avenues for its application in the biological sciences. [ABSTRACT FROM AUTHOR]

Endocytosis and relocalization of auxin carriers represent important mechanisms for adaptive plant growth and developmental responses. Both root gravitropism and halotropism have been shown to be dependent on relocalization of auxin transporters. Following their homology to mammalian phospholipase Ds (PLDs), plant PLDζ-type enzymes are likely candidates to regulate auxin carrier endocytosis. We investigated root tropic responses for an Arabidopsis pldζ1-KO mutant and its effect on the dynamics of two auxin transporters during salt stress, that is, PIN2 and AUX1. We found altered root growth and halotropic and gravitropic responses in the absence of PLDζ1 and report a role for PLDζ1 in the polar localization of PIN2. Additionally, irrespective of the genetic background, salt stress induced changes in AUX1 polarity. Utilizing our previous computational model, we found that these novel salt-induced AUX1 changes contribute to halotropic auxin asymmetry. We also report the formation of "osmotic stress-induced membrane structures." These large membrane structures are formed at the plasma membrane shortly after NaCl or sorbitol treatment and have a prolonged presence in a pldζ1 mutant. Taken together, these results show a crucial role for PLDζ1 in both ionic and osmotic stress-induced auxin carrier dynamics during salt stress., (© 2019 The Authors. Plant, Cell & Environment published by John Wiley & Sons Ltd.)

Somitogenesis is one of the major hallmarks of bilateral symmetry in vertebrates. This symmetry is lost when retinoic acid (RA) signalling is inhibited, allowing the left-right determination pathway to influence somitogenesis. In all three studied vertebrate model species, zebrafish, chicken and mouse, the frequency of somite formation becomes asymmetric, with slower gene expression oscillations driving somitogenesis on the right side. Still, intriguingly, the resulting left-right asymmetric phenotypes differ significantly between these model species. While somitogenesis is generally considered as functionally equivalent among different vertebrates, substantial differences exist in the subset of oscillating genes between different vertebrate species. Variation also appears to exist in the way oscillations cease and somite boundaries become patterned. In addition, in absence of RA, the FGF8 gradient thought to constitute the determination wavefront becomes asymmetric in zebrafish and mouse, extending more anteriorly to the right, while remaining symmetric in chicken. Here we use a computational modelling approach to decipher the causes underlying species differences in asymmetric somitogenesis. Specifically, we investigate to what extent differences can be explained from observed differences in FGF asymmetry and whether differences in somite determination dynamics may also be involved. We demonstrate that a simple clock-and-wavefront model incorporating the observed left-right differences in somitogenesis frequency readily reproduces asymmetric somitogenesis in chicken. However, incorporating asymmetry in FGF signalling was insufficient to robustly reproduce mouse or zebrafish asymmetry phenotypes. In order to explain these phenoptypes we needed to extend the basic model, incorporating species-specific details of the somitogenesis determination mechanism. Our results thus demonstrate that a combination of differences in FGF dynamics and somite determination cause species differences in asymmetric somitogenesis. In addition,they highlight the power of using computational models as well as studying left-right asymmetry to obtain more insight in somitogenesis., (Copyright © 2017 Elsevier Inc. All rights reserved.)

Background: Sexual reproduction has classically been considered as a barrier to the buildup of discrete phenotypic differentiation. This notion has been confirmed by models of sympatric speciation in which a fixed genetic architecture and a linear genotype phenotype mapping were assumed. In this paper we study the influence of a flexible genetic architecture and non-linear genotype phenotype map on differentiation under sexual reproduction.We use an individual based model in which organisms have a genome containing genes and transcription factor binding sites. Mutations involve single genes or binding sites or stretches of genome. The genome codes for a regulatory network that determines the gene expression pattern and hence the phenotype of the organism, resulting in a non-linear genotype phenotype map. The organisms compete in a multi-niche environment, imposing selection for phenotypic differentiation., Results: We find as a generic outcome the evolution of discrete clusters of organisms adapted to different niches, despite random mating. Organisms from different clusters are distinct on the genotypic, the network and the phenotypic level. However, the genome and network differences are constrained to a subset of the genome locations, a process we call genotypic canalization. We demonstrate how this canalization leads to an increased robustness to recombination and increasing hybrid fitness. Finally, in case of assortative mating, we explain how this canalization increases the effectiveness of assortativeness., Conclusion: We conclude that in case of a flexible genetic architecture and a non-linear genotype phenotype mapping, sexual reproduction does not constrain phenotypic differentiation, but instead constrains the genotypic differences underlying it. We hypothesize that, as genotypic canalization enables differentiation despite random mating and increases the effectiveness of assortative mating, sympatric speciation is more likely than is commonly suggested.

Aims: During ageing, after infarction, in cardiomyopathies and other cardiac diseases, the percentage of fibrotic (connective) tissue may increase from 6% up to 10-35%. The presence of increased amounts of connective tissue is strongly correlated with the occurrence of arrhythmias and sudden cardiac death., Methods and Results: In this article, we investigate the role of diffuse fibrosis on wave propagation, arrhythmogenesis, and arrhythmia mechanism in human ventricular tissue using computer modelling. We show that diffuse fibrosis slows down wave propagation and increases tissue vulnerability to wave break and spiral wave formation. We also demonstrate that diffuse fibrosis increases the period of re-entrant arrhythmias and can suppress the restitution-induced transition from tachycardia to fibrillation., Conclusion: The latter suggests that mechanisms different from restitution-induced spiral break-up might be more likely to account for the onset of fibrillation in the presence of large amounts of diffuse fibrotic tissue.

Sudden cardiac death is a major cause of death in the industrialized world, claiming approximately 300,000 victims annually in the United States alone. In most cases, sudden cardiac death is caused by ventricular fibrillation (VF). Experimental studies in large animal hearts have shown that the uncoordinated contractions during VF are caused by large numbers of chaotically wandering reentrant waves of electrical activity. However, recent clinical data on VF in the human heart seem to suggest that human VF may have a markedly different organization. Here, we use a detailed model of the human ventricles, including a detailed description of cell electrophysiology, ventricular anatomy, and fiber direction anisotropy, to study the organization of human VF. We show that characteristics of our simulated VF are qualitatively similar to the clinical data. Furthermore, we find that human VF is driven by only approximately 10 reentrant sources and thus is much more organized than VF in animal hearts of comparable size, where VF is driven by approximately 50 sources. We investigate the influence of anisotropy ratio, tissue excitability, and restitution properties on the number of reentrant sources driving VF. We find that the number of rotors depends strongest on minimum action potential duration, a property that differs significantly between human and large animal hearts. Based on these findings, we suggest that the simpler spatial organization of human VF relative to VF in large animal hearts may be caused by differences in minimum action potential duration. Both the simpler spatial organization of human VF and its suggested cause may have important implications for treating and preventing this dangerous arrhythmia in humans.

In this paper we briefly review currently published models for human ventricular cells and tissues. We discuss the Priebe-Beuckelmann (PB) model and the reduced version of this model constructed by Bernus et al. (redPB), the Ten Tusscher-Noble-Noble-Panfilov (TNNP) model and the Iyer-Mazhari-Winslow (IMW) model. We compare several characteristics of these models such as: sources of experimental data the models are based on, action potential morphology, action potential duration (APD) and conduction velocity (CV) restitution and computational efficiency. Finally, we discuss the application of a subset of these models-the redPB and the TNNP model-to study simulated spiral wave dynamics in 2D tissue sheets and in the human ventricles. We discuss the suitability of the different models for particular research questions and their limitations.