Search

Background: Homologous recombination deficiency (HRD) is recognized as a pan-cancer predictive biomarker that potentially indicates who could benefit from treatment with PARP inhibitors (PARPi). Despite its clinical significance, HRD testing is highly complex. Here, we investigated in a proof-of-concept study whether Deep Learning (DL) can predict HRD status solely based on routine hematoxylin & eosin (H&E) histology images across nine different cancer types., Methods: We developed a deep learning pipeline with attention-weighted multiple instance learning (attMIL) to predict HRD status from histology images. As part of our approach, we calculated a genomic scar HRD score by combining loss of heterozygosity (LOH), telomeric allelic imbalance (TAI), and large-scale state transitions (LST) from whole genome sequencing (WGS) data of n = 5209 patients across two independent cohorts. The model's effectiveness was evaluated using the area under the receiver operating characteristic curve (AUROC), focusing on its accuracy in predicting genomic HRD against a clinically recognized cutoff value., Results: Our study demonstrated the predictability of genomic HRD status in endometrial, pancreatic, and lung cancers reaching cross-validated AUROCs of 0.79, 0.58, and 0.66, respectively. These predictions generalized well to an external cohort, with AUROCs of 0.93, 0.81, and 0.73. Moreover, a breast cancer-trained image-based HRD classifier yielded an AUROC of 0.78 in the internal validation cohort and was able to predict HRD in endometrial, prostate, and pancreatic cancer with AUROCs of 0.87, 0.84, and 0.67, indicating that a shared HRD-like phenotype occurs across these tumor entities., Conclusions: This study establishes that HRD can be directly predicted from H&E slides using attMIL, demonstrating its applicability across nine different tumor types., (© 2024. The Author(s).)

Most clinical information is encoded as free text, not accessible for quantitative analysis. This study presents an open-source pipeline using the local large language model (LLM) "Llama 2" to extract quantitative information from clinical text and evaluates its performance in identifying features of decompensated liver cirrhosis. The LLM identified five key clinical features in a zero- and one-shot manner from 500 patient medical histories in the MIMIC IV dataset. We compared LLMs of three sizes and various prompt engineering approaches, with predictions compared against ground truth from three blinded medical experts. Our pipeline achieved high accuracy, detecting liver cirrhosis with 100% sensitivity and 96% specificity. High sensitivities and specificities were also yielded for detecting ascites (95%, 95%), confusion (76%, 94%), abdominal pain (84%, 97%), and shortness of breath (87%, 97%) using the 70 billion parameter model, which outperformed smaller versions. Our study successfully demonstrates the capability of locally deployed LLMs to extract clinical information from free text with low hardware requirements., (© 2024. The Author(s).)

Hematoxylin- and eosin-stained whole-slide images (WSIs) are the foundation of diagnosis of cancer. In recent years, development of deep learning-based methods in computational pathology has enabled the prediction of biomarkers directly from WSIs. However, accurately linking tissue phenotype to biomarkers at scale remains a crucial challenge for democratizing complex biomarkers in precision oncology. This protocol describes a practical workflow for solid tumor associative modeling in pathology (STAMP), enabling prediction of biomarkers directly from WSIs by using deep learning. The STAMP workflow is biomarker agnostic and allows for genetic and clinicopathologic tabular data to be included as an additional input, together with histopathology images. The protocol consists of five main stages that have been successfully applied to various research problems: formal problem definition, data preprocessing, modeling, evaluation and clinical translation. The STAMP workflow differentiates itself through its focus on serving as a collaborative framework that can be used by clinicians and engineers alike for setting up research projects in the field of computational pathology. As an example task, we applied STAMP to the prediction of microsatellite instability (MSI) status in colorectal cancer, showing accurate performance for the identification of tumors high in MSI. Moreover, we provide an open-source code base, which has been deployed at several hospitals across the globe to set up computational pathology workflows. The STAMP workflow requires one workday of hands-on computational execution and basic command line knowledge., (© 2024. Springer Nature Limited.)

In clinical science and practice, text data, such as clinical letters or procedure reports, is stored in an unstructured way. This type of data is not a quantifiable resource for any kind of quantitative investigations and any manual review or structured information retrieval is time-consuming and costly. The capabilities of Large Language Models (LLMs) mark a paradigm shift in natural language processing and offer new possibilities for structured Information Extraction (IE) from medical free text. This protocol describes a workflow for LLM based information extraction (LLM-AIx), enabling extraction of predefined entities from unstructured text using privacy preserving LLMs. By converting unstructured clinical text into structured data, LLM-AIx addresses a critical barrier in clinical research and practice, where the efficient extraction of information is essential for improving clinical decision-making, enhancing patient outcomes, and facilitating large-scale data analysis. The protocol consists of four main processing steps: 1) Problem definition and data preparation, 2) data preprocessing, 3) LLM-based IE and 4) output evaluation. LLM-AIx allows integration on local hospital hardware without the need of transferring any patient data to external servers. As example tasks, we applied LLM-AIx for the anonymization of fictitious clinical letters from patients with pulmonary embolism. Additionally, we extracted symptoms and laterality of the pulmonary embolism of these fictitious letters. We demonstrate troubleshooting for potential problems within the pipeline with an IE on a real-world dataset, 100 pathology reports from the Cancer Genome Atlas Program (TCGA), for TNM stage extraction. LLM-AIx can be executed without any programming knowledge via an easy-to-use interface and in no more than a few minutes or hours, depending on the LLM model selected., Competing Interests: JNK declares consulting services for Owkin, France; DoMore Diagnostics, Norway; Panakeia, UK; AstraZeneca, UK; Scailyte, Switzerland; Mindpeak, Germany; and MultiplexDx, Slovakia. Furthermore he holds shares in StratifAI GmbH, Germany, has received a research grant by GSK, and has received honoraria by AstraZeneca, Bayer, Eisai, Janssen, MSD, BMS, Roche, Pfizer and Fresenius. ICW has received honoraria by AstraZeneca. No further potential COIs are disclosed by any of the authors. KKB has received honoraria by Canon Medical Systems Corporations and GE Healthcare.

Most biological processes are regulated by signaling modules that bind to short linear motifs. For protein kinases, substrates may have full or only partial matches to the kinase recognition motif, a property known as "substrate quality." However, it is not clear whether differences in substrate quality represent neutral variation or if they have functional consequences. We examine this question for the kinase CK2, which has many fundamental functions. We show that optimal CK2 sites are phosphorylated at maximal stoichiometries and found in many conditions, whereas minimal substrates are more weakly phosphorylated and have regulatory functions. Optimal CK2 sites tend to be more conserved, and substrate quality is often tuned by selection. For intermediate sites, increases or decreases in substrate quality may be deleterious, as we demonstrate for a CK2 substrate at the kinetochore. The results together suggest a strong role for substrate quality in phosphosite function and evolution. A record of this paper's transparent peer review process is included in the supplemental information., Competing Interests: Declaration of interests The authors declare no competing interests., (Copyright © 2024 The Author(s). Published by Elsevier Inc. All rights reserved.)

Plasmodium falciparum, the deadly protozoan parasite responsible for malaria, has a tightly regulated gene expression profile closely linked to its intraerythrocytic development cycle. Epigenetic modifiers of the histone acetylation code have been identified as key regulators of the parasite's transcriptome but require further investigation. In this study, we map the genomic distribution of Plasmodium falciparum histone deacetylase 1 (PfHDAC1) across the erythrocytic asexual development cycle and find it has a dynamic occupancy over a wide array of developmentally relevant genes. Overexpression of PfHDAC1 results in a progressive increment in parasite load over consecutive rounds of the asexual infection cycle and is associated with enhanced gene expression of multiple families of host cell invasion factors (merozoite surface proteins, rhoptry proteins, etc.) and with increased merozoite invasion efficiency. With the use of class-specific inhibitors, we demonstrate that PfHDAC1 activity in parasites is crucial for timely intraerythrocytic development. Interestingly, overexpression of PfHDAC1 results in decreased sensitivity to frontline-drug dihydroartemisinin in parasites. Furthermore, we identify that artemisinin exposure can interfere with PfHDAC1 abundance and chromatin occupancy, resulting in enrichment over genes implicated in response/resistance to artemisinin. Finally, we identify that dihydroartemisinin exposure can interrupt the in vitro catalytic deacetylase activity and post-translational phosphorylation of PfHDAC1, aspects that are crucial for its genomic function. Collectively, our results demonstrate PfHDAC1 to be a regulator of critical functions in asexual parasite development and host invasion, which is responsive to artemisinin exposure stress and deterministic of resistance to it., Importance: Malaria is a major public health problem, with the parasite Plasmodium falciparum causing most of the malaria-associated mortality. It is spread by the bite of infected mosquitoes and results in symptoms such as cyclic fever, chills, and headache. However, if left untreated, it can quickly progress to a more severe and life-threatening form. The World Health Organization currently recommends the use of artemisinin combination therapy, and it has worked as a gold standard for many years. Unfortunately, certain countries in southeast Asia and Africa, burdened with a high prevalence of malaria, have reported cases of drug-resistant infections. One of the major problems in controlling malaria is the emergence of artemisinin resistance. Population genomic studies have identified mutations in the Kelch13 gene as a molecular marker for artemisinin resistance. However, several reports thereafter indicated that Kelch13 is not the main mediator but rather hinted at transcriptional deregulation as a major determinant of drug resistance. Earlier, we identified PfGCN5 as a global regulator of stress-responsive genes, which are known to play a central role in artemisinin resistance generation. In this study, we have identified PfHDAC1, a histone deacetylase as a cell cycle regulator, playing an important role in artemisinin resistance generation. Taken together, our study identified key transcriptional regulators that play an important role in artemisinin resistance generation., Competing Interests: The authors declare no conflict of interest.

In the spectrum of colorectal tumors, microsatellite-stable (MSS) tumors with DNA polymerase ε (POLE) mutations exhibit a hypermutated profile, holding the potential to respond to immunotherapy similarly to their microsatellite-instable (MSI) counterparts. Yet, due to their rarity and the associated testing costs, systematic screening for these mutations is not commonly pursued. Notably, the histopathological phenotype resulting from POLE mutations is theorized to resemble that of MSI. This resemblance not only could facilitate their detection by a transformer-based Deep Learning (DL) system trained on MSI pathology slides, but also indicates the possibility for MSS patients with POLE mutations to access enhanced treatment options, which might otherwise be overlooked. To harness this potential, we trained a Deep Learning classifier on a large dataset with the ground truth for microsatellite status and subsequently validated its capabilities for MSI and POLE detection across three external cohorts. Our model accurately identified MSI status in both the internal and external resection cohorts using pathology images alone. Notably, with a classification threshold of 0.5, over 75% of POLE driver mutant patients in the external resection cohorts were flagged as "positive" by a DL system trained on MSI status. In a clinical setting, deploying this DL model as a preliminary screening tool could facilitate the efficient identification of clinically relevant MSI and POLE mutations in colorectal tumors, in one go., (© 2024. The Author(s).)

Background: Sporozoite invasion of hepatocytes is an essential step in the Plasmodium life-cycle and has similarities, at the cellular level, to merozoite invasion of erythrocytes. In the case of the Plasmodium blood-stage, efforts to identify host-pathogen protein-protein interactions have yielded important insights including vaccine candidates. In the case of sporozoite-hepatocyte invasion, the host-pathogen protein-protein interactions involved are poorly understood., Methods: To gain a better understanding of the protein-protein interaction between the sporozoite ligands and host receptors, a systematic screen was performed. The previous Plasmodium falciparum and human surface protein ectodomain libraries were substantially extended, resulting in the creation of new libraries comprising 88 P. falciparum sporozoite protein coding sequences and 182 sequences encoding human hepatocyte surface proteins. Having expressed recombinant proteins from these sequences, a plate-based assay was used, capable of detecting low affinity interactions between recombinant proteins, modified for enhanced throughput, to screen the proteins for interactions. The novel interactions identified in the screen were characterized biochemically, and their essential role in parasite invasion was further elucidated using antibodies and genetically manipulated Plasmodium parasites., Results: A total of 7540 sporozoite-hepatocyte protein pairs were tested under conditions capable of detecting interactions of at least 1.2 µM K D . An interaction between the human fibroblast growth factor receptor 4 (FGFR4) and the P. falciparum protein Pf34 is identified and reported here, characterizing its affinity and demonstrating the blockade of the interaction by reagents, including a monoclonal antibody. Furthermore, further interactions between Pf34 and a second P. falciparum rhoptry neck protein, PfRON6, and between human low-density lipoprotein receptor (LDLR) and the P. falciparum protein PIESP15 are identified. Conditional genetic deletion confirmed the essentiality of PfRON6 in the blood-stage, consistent with the important role of this protein in parasite lifecycle. Pf34 was refractory to attempted genetic modification. Antibodies to Pf34 abrogated the interaction and had a modest effect upon sporozoite invasion into primary human hepatocytes., Conclusion: Pf34 and PfRON6 may be members of a functionally important invasion complex which could be a target for future interventions. The modified interaction screening assay, protein expression libraries and P. falciparum mutant parasites reported here may be a useful tool for protein interaction discovery and antigen candidate screening which could be of wider value to the scientific community., (© 2024. The Author(s).)

The Plasmodium falciparum merozoite surface protein MSPDBL2 is a polymorphic antigen targeted by acquired immune responses, and normally expressed in only a minority of mature schizonts. The potential relationship of MSPDBL2 to sexual commitment is examined, as variable mspdbl2 transcript levels and proportions of MSPDBL2-positive mature schizonts in clinical isolates have previously correlated with levels of many sexual stage parasite gene transcripts, although not with the master regulator ap2-g . It is demonstrated that conditional overexpression of the gametocyte development protein GDV1, which promotes sexual commitment, also substantially increases the proportion of MSPDBL2-positive schizonts in culture. Conversely, truncation of the gdv1 gene is shown to prevent any expression of MSPDBL2. However, across diverse P. falciparum cultured lines, the variable proportions of MSPDBL2 positivity in schizonts do not correlate significantly with variable gametocyte conversion rates, indicating it is not involved in sexual commitment. Confirming this, examining a line with endogenous hemagglutinin-tagged AP2-G showed that the individual schizonts expressing MSPDBL2 are mostly different from those expressing AP2-G. Using a selection-linked integration system, modified P. falciparum lines were engineered to express an intact or disrupted version of MSPDBL2, showing the protein is not required for sexual commitment or early gametocyte development. Asexual parasite multiplication rates were also not affected by expression of either intact or disrupted MSPDBL2 in a majority of schizonts. Occurring alongside sexual commitment, the role of the discrete MSPDBL2-positive schizont subpopulation requires further investigation in natural infections where it is under immune selection., Importance: Malaria parasites in the blood are remarkably variable, able to switch antigenic targets so they may survive within humans who have already developed specific immune responses. This is one of the challenges in developing vaccines against malaria. MSPDBL2 is a target of naturally acquired immunity expressed in minority proportions of schizonts, the end stages of each 2-day replication cycle in red blood cells which contain merozoites prepared to invade new red blood cells. Results show that the proportion of schizonts expressing MSPDBL2 is positively controlled by the expression of the regulatory gametocyte development protein GDV1. It was previously known that expression of GDV1 leads to increased expression of AP2-G which causes parasites to switch to sexual development, so a surprising finding here is that MSPDBL2-positive parasites are mostly distinct from those that express AP2-G. This discrete antigenic subpopulation of mostly asexual parasites is regulated alongside sexually committed parasites, potentially enabling survival under stress conditions., Competing Interests: The authors declare no conflict of interest.

Deep Learning (DL) can predict biomarkers from cancer histopathology. Several clinically approved applications use this technology. Most approaches, however, predict categorical labels, whereas biomarkers are often continuous measurements. We hypothesize that regression-based DL outperforms classification-based DL. Therefore, we develop and evaluate a self-supervised attention-based weakly supervised regression method that predicts continuous biomarkers directly from 11,671 images of patients across nine cancer types. We test our method for multiple clinically and biologically relevant biomarkers: homologous recombination deficiency score, a clinically used pan-cancer biomarker, as well as markers of key biological processes in the tumor microenvironment. Using regression significantly enhances the accuracy of biomarker prediction, while also improving the predictions' correspondence to regions of known clinical relevance over classification. In a large cohort of colorectal cancer patients, regression-based prediction scores provide a higher prognostic value than classification-based scores. Our open-source regression approach offers a promising alternative for continuous biomarker analysis in computational pathology., (© 2024. The Author(s).)

Breast cancer prognosis and management for both men and women are reliant upon estrogen receptor alpha (ERα) and progesterone receptor (PR) expression to inform therapy. Previous studies have shown that there are sex-specific binding characteristics of ERα and PR in breast cancer and, counterintuitively, ERα expression is more common in male than female breast cancer. We hypothesized that these differences could have morphological manifestations that are undetectable to human observers but could be elucidated computationally. To investigate this, we trained attention-based multiple instance learning prediction models for ERα and PR using H&E-stained images of female breast cancer from the Cancer Genome Atlas (TCGA) (n = 1085) and deployed them on external female (n = 192) and male breast cancer images (n = 245). Both targets were predicted in the internal (AUROC for ERα prediction: 0.86 ± 0.02, p < 0.001; AUROC for PR prediction = 0.76 ± 0.03, p < 0.001) and external female cohorts (AUROC for ERα prediction: 0.78 ± 0.03, p < 0.001; AUROC for PR prediction = 0.80 ± 0.04, p < 0.001) but not the male cohort (AUROC for ERα prediction: 0.66 ± 0.14, p = 0.43; AUROC for PR prediction = 0.63 ± 0.04, p = 0.05). This suggests that subtle morphological differences invisible upon visual inspection may exist between the sexes, supporting previous immunohistochemical, genomic, and transcriptomic analyses., (© 2023. The Author(s).)

Apicomplexan parasites use Ca 2+ -regulated exocytosis to secrete essential virulence factors from specialized organelles called micronemes. Ca 2+ -dependent protein kinases (CDPKs) are required for microneme exocytosis; however, the molecular events that regulate trafficking and fusion of micronemes with the plasma membrane remain unresolved. Here, we combine sub-minute resolution phosphoproteomics and bio-orthogonal labeling of kinase substrates in Toxoplasma gondii to identify 163 proteins phosphorylated in a CDPK1-dependent manner. In addition to known regulators of secretion, we identify uncharacterized targets with predicted functions across signaling, gene expression, trafficking, metabolism, and ion homeostasis. One of the CDPK1 targets is a putative HOOK activating adaptor. In other eukaryotes, HOOK homologs form the FHF complex with FTS and FHIP to activate dynein-mediated trafficking of endosomes along microtubules. We show the FHF complex is partially conserved in T. gondii , consisting of HOOK, an FTS homolog, and two parasite-specific proteins (TGGT1&#95;306920 and TGGT1&#95;316650). CDPK1 kinase activity and HOOK are required for the rapid apical trafficking of micronemes as parasites initiate motility. Moreover, parasites lacking HOOK or FTS display impaired microneme protein secretion, leading to a block in the invasion of host cells. Taken together, our work provides a comprehensive catalog of CDPK1 targets and reveals how vesicular trafficking has been tuned to support a parasitic lifestyle., Competing Interests: AC, MB, EY, NH, SC, EA, AH, ES, MT No competing interests declared, SL Reviewing editor, eLife, (© 2023, Chan et al.)

Background: Deep Learning (DL) can predict molecular alterations of solid tumors directly from routine histopathology slides. Since the 2021 update of the World Health Organization (WHO) diagnostic criteria, the classification of brain tumors integrates both histopathological and molecular information. We hypothesize that DL can predict molecular alterations as well as WHO subtyping of brain tumors from hematoxylin and eosin-stained histopathology slides., Methods: We used weakly supervised DL and applied it to three large cohorts of brain tumor samples, comprising N  = 2845 patients., Results: We found that the key molecular alterations for subtyping, IDH and ATRX , as well as 1p19q codeletion, were predictable from histology with an area under the receiver operating characteristic curve (AUROC) of 0.95, 0.90, and 0.80 in the training cohort, respectively. These findings were upheld in external validation cohorts with AUROCs of 0.90, 0.79, and 0.87 for prediction of IDH , ATRX , and 1p19q codeletion, respectively., Conclusions: In the future, such DL-based implementations could ease diagnostic workflows, particularly for situations in which advanced molecular testing is not readily available., Competing Interests: J.N.K. declares consulting services for Owkin, France; DoMore Diagnostics, Norway and Panakeia, UK; furthermore, he holds shares in StratifAI GmbH and has received honoraria for lectures by AstraZeneca, Bayer, Eisai, MSD, BMS, Roche, Pfizer, and Fresenius. No other potential conflict of interest are noted by any of the authors., (© The Author(s) 2023. Published by Oxford University Press, the Society for Neuro-Oncology and the European Association of Neuro-Oncology.)

Intracellular pathogens and other endosymbionts reprogram host cell transcription to suppress immune responses and recalibrate biosynthetic pathways. This reprogramming is critical in determining the outcome of infection or colonization. We combine pooled CRISPR knockout screening with dual host-microbe single-cell RNA sequencing, a method we term dual perturb-seq, to identify the molecular mediators of these transcriptional interactions. Applying dual perturb-seq to the intracellular pathogen Toxoplasma gondii, we are able to identify previously uncharacterized effector proteins and directly infer their function from the transcriptomic data. We show that TgGRA59 contributes to the export of other effector proteins from the parasite into the host cell and identify an effector, TgSOS1, that is necessary for sustained host STAT6 signaling and thereby contributes to parasite immune evasion and persistence. Together, this work demonstrates a tool that can be broadly adapted to interrogate host-microbe transcriptional interactions and reveal mechanisms of infection and immune evasion., Competing Interests: Declaration of interests The authors declare no competing interests., (Copyright © 2023 The Author(s). Published by Elsevier Inc. All rights reserved.)

Reverse genetics is key to understanding protein function, but the mechanistic connection between a gene of interest and the observed phenotype is not always clear. Here we describe the use of proximity labeling using TurboID and site-specific quantification of biotinylated peptides to measure changes to the local protein environment of selected targets upon perturbation. We apply this technique, which we call PerTurboID, to understand how the Plasmodium falciparum -exported kinase, FIKK4.1, regulates the function of the major virulence factor of the malaria-causing parasite, PfEMP1. We generated independent TurboID fusions of two proteins that are predicted substrates of FIKK4.1 in a FIKK4.1 conditional KO parasite line. Comparing the abundance of site-specific biotinylated peptides between wildtype and kinase deletion lines reveals the differential accessibility of proteins to biotinylation, indicating changes to localization, protein-protein interactions, or protein structure which are mediated by FIKK4.1 activity. We further show that FIKK4.1 is likely the only FIKK kinase that controls surface levels of PfEMP1, but not other surface antigens, on the infected red blood cell under standard culture conditions. We believe PerTurboID is broadly applicable to study the impact of genetic or environmental perturbation on a selected cellular niche., Competing Interests: HD, HB, MB, JD, MT No competing interests declared, (© 2023, Davies et al.)

Malaria parasites in the blood stage express a single transmembrane transport protein for the release of the glycolytic end product l-lactate/H + from the cell. This transporter is a member of the strictly microbial formate-nitrite transporter (FNT) family and a novel putative drug target. Small, drug-like FNT inhibitors potently block lactate transport and kill Plasmodium falciparum parasites in culture. The protein structure of Plasmodium falciparum FNT (PfFNT) in complex with the inhibitor has been resolved and confirms its previously predicted binding site and its mode of action as a substrate analog. Here, we investigated the mutational plasticity and essentiality of the PfFNT target on a genetic level, and established its in vivo druggability using mouse malaria models. We found that, besides a previously identified PfFNT G107S resistance mutation, selection of parasites at 3 × IC 50 (50% inhibitory concentration) gave rise to two new point mutations affecting inhibitor binding: G21E and V196L. Conditional knockout and mutation of the PfFNT gene showed essentiality in the blood stage, whereas no phenotypic defects in sexual development were observed. PfFNT inhibitors mainly targeted the trophozoite stage and exhibited high potency in P. berghei- and P. falciparum-infected mice. Their in vivo activity profiles were comparable to that of artesunate, demonstrating strong potential for the further development of PfFNT inhibitors as novel antimalarials., Competing Interests: The authors declare no conflict of interest.

Toxoplasma gondii secretes protein effectors to subvert the human immune system sufficiently to establish a chronic infection. Relative to murine infections, little is known about which parasite effectors disarm human immune responses. Here, we used targeted CRISPR screening to identify secreted protein effectors required for parasite survival in IFNγ-activated human cells. Independent screens were carried out using 2 Toxoplasma strains that differ in virulence in mice, leading to the identification of effectors required for survival in IFNγ-activated human cells. We identify the secreted protein GRA57 and 2 other proteins, GRA70 and GRA71, that together form a complex which enhances the ability of parasites to persist in IFNγ-activated human foreskin fibroblasts (HFFs). Components of the protein machinery required for export of Toxoplasma proteins into the host cell were also found to be important for parasite resistance to IFNγ in human cells, but these export components function independently of the identified protein complex. Host-mediated ubiquitination of the parasite vacuole has previously been associated with increased parasite clearance from human cells, but we find that vacuoles from GRA57, GRA70, and GRA71 knockout strains are surprisingly less ubiquitinated by the host cell. We hypothesise that this is likely a secondary consequence of deletion of the complex, unlinked to the IFNγ resistance mediated by these effectors., Competing Interests: The authors have declared that no competing interests exist., (Copyright: © 2023 Lockyer et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.)

Deep learning (DL) can predict microsatellite instability (MSI) from routine histopathology slides of colorectal cancer (CRC). However, it is unclear whether DL can also predict other biomarkers with high performance and whether DL predictions generalize to external patient populations. Here, we acquire CRC tissue samples from two large multi-centric studies. We systematically compare six different state-of-the-art DL architectures to predict biomarkers from pathology slides, including MSI and mutations in BRAF, KRAS, NRAS, and PIK3CA. Using a large external validation cohort to provide a realistic evaluation setting, we show that models using self-supervised, attention-based multiple-instance learning consistently outperform previous approaches while offering explainable visualizations of the indicative regions and morphologies. While the prediction of MSI and BRAF mutations reaches a clinical-grade performance, mutation prediction of PIK3CA, KRAS, and NRAS was clinically insufficient., Competing Interests: Declaration of interests For transparency, we provide the following information: J.N.K. declares consulting services for Owkin, France; Panakeia, UK; and DoMore Diagnostics, Norway. P.Q. and N.P.W. declare research funding from Roche and PQ consulting and speaker services for Roche. P.Q. is a National Institute of Health Research senior investigator., (Copyright © 2023 The Author(s). Published by Elsevier Inc. All rights reserved.)

The histopathological phenotype of tumors reflects the underlying genetic makeup. Deep learning can predict genetic alterations from pathology slides, but it is unclear how well these predictions generalize to external datasets. We performed a systematic study on Deep-Learning-based prediction of genetic alterations from histology, using two large datasets of multiple tumor types. We show that an analysis pipeline that integrates self-supervised feature extraction and attention-based multiple instance learning achieves a robust predictability and generalizability., (© 2023. The Author(s).)

Aims: One of the most important complications of heart transplantation is organ rejection, which is diagnosed on endomyocardial biopsies by pathologists. Computer-based systems could assist in the diagnostic process and potentially improve reproducibility. Here, we evaluated the feasibility of using deep learning in predicting the degree of cellular rejection from pathology slides as defined by the International Society for Heart and Lung Transplantation (ISHLT) grading system., Methods and Results: We collected 1079 histopathology slides from 325 patients from three transplant centres in Germany. We trained an attention-based deep neural network to predict rejection in the primary cohort and evaluated its performance using cross-validation and by deploying it to three cohorts. For binary prediction (rejection yes/no), the mean area under the receiver operating curve (AUROC) was 0.849 in the cross-validated experiment and 0.734, 0.729, and 0.716 in external validation cohorts. For a prediction of the ISHLT grade (0R, 1R, 2/3R), AUROCs were 0.835, 0.633, and 0.905 in the cross-validated experiment and 0.764, 0.597, and 0.913; 0.631, 0.633, and 0.682; and 0.722, 0.601, and 0.805 in the validation cohorts, respectively. The predictions of the artificial intelligence model were interpretable by human experts and highlighted plausible morphological patterns., Conclusion: We conclude that artificial intelligence can detect patterns of cellular transplant rejection in routine pathology, even when trained on small cohorts., Competing Interests: Conflict of interest: J.N.K. declares consulting services for Owkin, France, and Panakeia, UK, as well as reimbursement for scientific talks by MSD, Eisai, and Fresenius. D.W. declares consulting services and honorary talks for Abiomed, AstraZeneca, Bayer, Berlin-Chemie, Novartis, Medtronic. CM declares honorary talks for AstraZeneca, Novartis, Heinen&Loewenstein, Boehringer Ingelheim/Lilly, Bayer, Pfizer, Sanofi, Aventis, Apontis, and Abbott and meeting support from AstraZeneca, Novartis, and Boehringer Ingelheim/Lilly. T.L. declares consulting fees from AstraZeneca, BMS, EISAI, Incyte, MSD, Roche, and HepaRegeniX and honorary talks and travel support from Abbvie and Gilead. The other authors do not have anything to disclose., (© The Author(s) 2023. Published by Oxford University Press on behalf of the European Society of Cardiology.)

Background: Computational pathology uses deep learning (DL) to extract biomarkers from routine pathology slides. Large multicentric datasets improve performance, but such datasets are scarce for gastric cancer. This limitation could be overcome by Swarm Learning (SL)., Methods: Here, we report the results of a multicentric retrospective study of SL for prediction of molecular biomarkers in gastric cancer. We collected tissue samples with known microsatellite instability (MSI) and Epstein-Barr Virus (EBV) status from four patient cohorts from Switzerland, Germany, the UK and the USA, storing each dataset on a physically separate computer., Results: On an external validation cohort, the SL-based classifier reached an area under the receiver operating curve (AUROC) of 0.8092 (± 0.0132) for MSI prediction and 0.8372 (± 0.0179) for EBV prediction. The centralized model, which was trained on all datasets on a single computer, reached a similar performance., Conclusions: Our findings demonstrate the feasibility of SL-based molecular biomarkers in gastric cancer. In the future, SL could be used for collaborative training and, thus, improve the performance of these biomarkers. This may ultimately result in clinical-grade performance and generalizability., (© 2022. The Author(s).)

Toxoplasma gondii is an intracellular parasite that can infect many host species and is a cause of significant human morbidity worldwide. T. gondii secretes a diverse array of effector proteins into the host cell which are critical for infection. The vast majority of these secreted proteins have no predicted functional domains and remain uncharacterised. Here, we carried out a pooled CRISPR knockout screen in the T. gondii Prugniaud strain in vivo to identify secreted proteins that contribute to parasite immune evasion in the host. We demonstrate that ROP1, the first-identified rhoptry protein of T. gondii, is essential for virulence and has a previously unrecognised role in parasite resistance to interferon gamma-mediated innate immune restriction. This function is conserved in the highly virulent RH strain of T. gondii and contributes to parasite growth in both murine and human macrophages. While ROP1 affects the morphology of rhoptries, from where the protein is secreted, it does not affect rhoptry secretion. Finally, we show that ROP1 co-immunoprecipitates with the host cell protein C1QBP, an emerging regulator of innate immune signaling. In summary, we identify putative in vivo virulence factors in the T. gondii Prugniaud strain and show that ROP1 is an important and previously overlooked effector protein that counteracts both murine and human innate immunity., Competing Interests: The authors have declared that no competing interests exist., (Copyright: © 2022 Butterworth et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.)

Fundamental processes that govern the lytic cycle of the intracellular parasite Toxoplasma gondii are regulated by several signalling pathways. However, how these pathways are connected remains largely unknown. Here, we compare the phospho-signalling networks during Toxoplasma egress from its host cell by artificially raising cGMP or calcium levels. We show that both egress inducers trigger indistinguishable signalling responses and provide evidence for a positive feedback loop linking calcium and cyclic nucleotide signalling. Using WT and conditional knockout parasites of the non-essential calcium-dependent protein kinase 3 (CDPK3), which display a delay in calcium inonophore-mediated egress, we explore changes in phosphorylation and lipid signalling in sub-minute timecourses after inducing Ca2+ release. These studies indicate that cAMP and lipid metabolism are central to the feedback loop, which is partly dependent on CDPK3 and allows the parasite to respond faster to inducers of egress. Biochemical analysis of 4 phosphodiesterases (PDEs) identified in our phosphoproteomes establishes PDE2 as a cAMP-specific PDE which regulates Ca2+ induced egress in a CDPK3-independent manner. The other PDEs display dual hydrolytic activity and play no role in Ca2+ induced egress. In summary, we uncover a positive feedback loop that enhances signalling during egress, thereby linking several signalling pathways., Competing Interests: The authors have declared that no competing interests exist.

Artificial Intelligence (AI) can support diagnostic workflows in oncology by aiding diagnosis and providing biomarkers directly from routine pathology slides. However, AI applications are vulnerable to adversarial attacks. Hence, it is essential to quantify and mitigate this risk before widespread clinical use. Here, we show that convolutional neural networks (CNNs) are highly susceptible to white- and black-box adversarial attacks in clinically relevant weakly-supervised classification tasks. Adversarially robust training and dual batch normalization (DBN) are possible mitigation strategies but require precise knowledge of the type of attack used in the inference. We demonstrate that vision transformers (ViTs) perform equally well compared to CNNs at baseline, but are orders of magnitude more robust to white- and black-box attacks. At a mechanistic level, we show that this is associated with a more robust latent representation of clinically relevant categories in ViTs compared to CNNs. Our results are in line with previous theoretical studies and provide empirical evidence that ViTs are robust learners in computational pathology. This implies that large-scale rollout of AI models in computational pathology should rely on ViTs rather than CNN-based classifiers to provide inherent protection against perturbation of the input data, especially adversarial attacks., (© 2022. The Author(s).)

Artificial intelligence (AI) can predict the presence of molecular alterations directly from routine histopathology slides. However, training robust AI systems requires large datasets for which data collection faces practical, ethical and legal obstacles. These obstacles could be overcome with swarm learning (SL), in which partners jointly train AI models while avoiding data transfer and monopolistic data governance. Here, we demonstrate the successful use of SL in large, multicentric datasets of gigapixel histopathology images from over 5,000 patients. We show that AI models trained using SL can predict BRAF mutational status and microsatellite instability directly from hematoxylin and eosin (H&E)-stained pathology slides of colorectal cancer. We trained AI models on three patient cohorts from Northern Ireland, Germany and the United States, and validated the prediction performance in two independent datasets from the United Kingdom. Our data show that SL-trained AI models outperform most locally trained models, and perform on par with models that are trained on the merged datasets. In addition, we show that SL-based AI models are data efficient. In the future, SL can be used to train distributed AI models for any histopathology image analysis task, eliminating the need for data transfer., (© 2022. The Author(s).)

Artificial intelligence (AI) is widely used to analyze gastrointestinal (GI) endoscopy image data. AI has led to several clinically approved algorithms for polyp detection, but application of AI beyond this specific task is limited by the high cost of manual annotations. Here, we show that a weakly supervised AI can be trained on data from a clinical routine database to learn visual patterns of GI diseases without any manual labeling or annotation. We trained a deep neural network on a dataset of N = 29,506 gastroscopy and N = 18,942 colonoscopy examinations from a large endoscopy unit serving patients in Germany, the Netherlands and Belgium, using only routine diagnosis data for the 42 most common diseases. Despite a high data heterogeneity, the AI system reached a high performance for diagnosis of multiple diseases, including inflammatory, degenerative, infectious and neoplastic diseases. Specifically, a cross-validated area under the receiver operating curve (AUROC) of above 0.70 was reached for 13 diseases, and an AUROC of above 0.80 was reached for two diseases in the primary data set. In an external validation set including six disease categories, the AI system was able to significantly predict the presence of diverticulosis, candidiasis, colon and rectal cancer with AUROCs above 0.76. Reverse engineering the predictions demonstrated that plausible patterns were learned on the level of images and within images and potential confounders were identified. In summary, our study demonstrates the potential of weakly supervised AI to generate high-performing classifiers and identify clinically relevant visual patterns based on non-annotated routine image data in GI endoscopy and potentially other clinical imaging modalities., (© 2022. The Author(s).)

In the last four years, advances in Deep Learning technology have enabled the inference of selected mutational alterations directly from routine histopathology slides. In particular, recent studies have shown that genetic changes in clinically relevant driver genes are reflected in the histological phenotype of solid tumors and can be inferred by analysing routine Haematoxylin and Eosin (H&E) stained tissue sections with Deep Learning. However, these studies mostly focused on selected individual genes in selected tumor types. In addition, genetic changes in solid tumors primarily act by changing signaling pathways that regulate cell behaviour. In this study, we hypothesized that Deep Learning networks can be trained to directly predict alterations of genes and pathways across a spectrum of solid tumors. We manually outlined tumor tissue in H&E-stained tissue sections from 7,829 patients with 23 different tumor types from The Cancer Genome Atlas. We then trained convolutional neural networks in an end-to-end way to detect alterations in the most clinically relevant pathways or genes, directly from histology images. Using this automatic approach, we found that alterations in 12 out of 14 clinically relevant pathways and numerous single gene alterations appear to be detectable in tissue sections, many of which have not been reported before. Interestingly, we show that the prediction performance for single gene alterations is better than that for pathway alterations. Collectively, these data demonstrate the predictability of genetic alterations directly from routine cancer histology images and show that individual genes leave a stronger morphological signature than genetic pathways., Competing Interests: JK declares consulting services for Owkin, France and Panakeia, UK. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2022 Loeffler, Gaisa, Muti, van Treeck, Echle, Ghaffari Laleh, Trautwein, Heij, Grabsch, Ortiz Bruechle and Kather.)

Plasmodium malaria parasites are obligate intracellular protozoans that use a unique form of locomotion, termed gliding motility, to move through host tissues and invade cells. The process is substrate dependent and powered by an actomyosin motor that drives the posterior translocation of extracellular adhesins which, in turn, propel the parasite forward. Gliding motility is essential for tissue translocation in the sporozoite and ookinete stages; however, the short-lived erythrocyte-invading merozoite stage has never been observed to undergo gliding movement. Here we show Plasmodium merozoites possess the ability to undergo gliding motility in vitro and that this mechanism is likely an important precursor step for successful parasite invasion. We demonstrate that two human infective species, Plasmodium falciparum and Plasmodium knowlesi , have distinct merozoite motility profiles which may reflect distinct invasion strategies. Additionally, we develop and validate a higher throughput assay to evaluate the effects of genetic and pharmacological perturbations on both the molecular motor and the complex signaling cascade that regulates motility in merozoites. The discovery of merozoite motility provides a model to study the glideosome and adds a dimension for work aiming to develop treatments targeting the blood stage invasion pathways., Competing Interests: The authors declare no competing interest., (Copyright © 2021 the Author(s). Published by PNAS.)

Malaria is a mosquito-borne disease caused by apicomplexan parasites of the genus Plasmodium. Completion of the parasite's life cycle depends on the transmission of sexual stages, the gametocytes, from an infected human host to the mosquito vector. Sexual commitment occurs in only a small fraction of asexual blood-stage parasites and is initiated by external cues. The gametocyte development protein 1 (GDV1) has been described as a key facilitator to trigger sexual commitment. GDV1 interacts with the silencing factor heterochromatin protein 1 (HP1), leading to its dissociation from heterochromatic DNA at the genomic locus encoding AP2-G, the master transcription factor of gametocytogenesis. How this process is regulated is not known. In this study, we have addressed the role of protein kinases implicated in gametocyte development. From a pool of available protein kinase knockout (KO) lines, we identified two kinase knockout lines which fail to produce gametocytes. However, independent genetic verification revealed that both kinases are not required for gametocytogenesis but that both lines harbor the same mutation that leads to a truncation in the extreme C terminus of GDV1. Introduction of the identified nonsense mutation into the genome of wild-type parasite lines replicates the observed phenotype. Using a GDV1 overexpression line, we show that the truncation in the GDV1 C terminus does not interfere with the nuclear import of GDV1 or its interaction with HP1 in vitro but appears to be important to sustain GDV1 protein levels and thereby sexual commitment. IMPORTANCE Transmission of malaria-causing Plasmodium species by mosquitos requires the parasite to change from a continuously growing asexual parasite form growing in the blood to a sexually differentiated form, the gametocyte. Only a small subset of asexual parasites differentiates into gametocytes that are taken up by the mosquito. Transmission represents a bottleneck in the life cycle of the parasite, so a molecular understanding of the events that lead to stage conversion may identify novel intervention points. Here, we screened a subset of kinases we hypothesized to play a role in this process. While we did not identify kinases required for sexual conversion, we identified a mutation in the C terminus of the gametocyte development 1 protein (GDV1), which abrogates sexual development. The mutation destabilizes the protein but not its interaction with its cognate binding partner HP1. This suggests an important role for the GDV1 C terminus beyond trafficking and protein stability., (Copyright © 2021 Tibúrcio et al.)

The intracellular parasite Toxoplasma gondii resides within a membrane-bound parasitophorous vacuole (PV) and secretes an array of proteins to establish this replicative niche. It has been shown previously that Toxoplasma secretes kinases and that numerous proteins are phosphorylated after secretion. Here, we assess the role of the phosphorylation of strand-forming protein 1 (SFP1) and the related protein GRA29, two secreted proteins with unknown function. We show that both proteins form stranded structures in the PV that are independent of the previously described intravacuolar network or actin. SFP1 and GRA29 can each form these structures independently of other Toxoplasma secreted proteins, although GRA29 appears to regulate SFP1 strands. We show that an unstructured region at the C termini of SFP1 and GRA29 is required for the formation of strands and that mimicking the phosphorylation of this domain of SFP1 negatively regulates strand development. When tachyzoites convert to chronic-stage bradyzoites, both proteins show a dispersed localization throughout the cyst matrix. Many secreted proteins are reported to dynamically redistribute as the cyst forms, and secreted kinases are known to play a role in cyst formation. Using quantitative phosphoproteome and proteome analyses comparing tachyzoite and early bradyzoite stages, we reveal widespread differential phosphorylation of secreted proteins. While we found no direct evidence for phosphorylation playing a dominant role for SFP1/GRA29 redistribution in the cyst, these data support a model in which secreted kinases and phosphatases contribute to the regulation of secreted proteins during stage conversion. IMPORTANCE Toxoplasma gondii is a common parasite that infects up to one-third of the human population. Initially, the parasite grows rapidly, infecting and destroying cells of the host, but subsequently switches to a slow-growing form and establishes chronic infection. In both stages, the parasite lives within a membrane-bound vacuole within the host cell, but in the chronic stage, a durable cyst wall is synthesized, which provides protection to the parasite during transmission to a new host. Toxoplasma secretes proteins into the vacuole to build its replicative niche, and previous studies identified many of these proteins as phosphorylated. We investigate two secreted proteins and show that a phosphorylated region plays an important role in their regulation in acute stages. We also observed widespread phosphorylation of secreted proteins when parasites convert from acute to chronic stages, providing new insight into how the cyst wall may be dynamically regulated., (© Crown copyright 2020.)

N -myristoylation is a ubiquitous class of protein lipidation across eukaryotes and N -myristoyl transferase (NMT) has been proposed as an attractive drug target in several pathogens. Myristoylation often primes for subsequent palmitoylation and stable membrane attachment, however, growing evidence suggests additional regulatory roles for myristoylation on proteins. Here we describe the myristoylated proteome of Toxoplasma gondii using chemoproteomic methods and show that a small-molecule NMT inhibitor developed against related Plasmodium spp . is also functional in Toxoplasma . We identify myristoylation on a transmembrane protein, the microneme protein 7 (MIC7), which enters the secretory pathway in an unconventional fashion with the myristoylated N-terminus facing the lumen of the micronemes. MIC7 and its myristoylation play a crucial role in the initial steps of invasion, likely during the interaction with and penetration of the host cell. Myristoylation of secreted eukaryotic proteins represents a substantial expansion of the functional repertoire of this co-translational modification., Competing Interests: MB, CD, LV, SN, EB, BT, AH, BW, SF, SM, JY, IT, MT No competing interests declared, ET EWT is a founder, shareholder and Director of Myricx Pharma Ltd, (© 2020, Broncel et al.)

The most severe form of human malaria is caused by Plasmodium falciparum. Its virulence is closely linked to the increase in rigidity of infected erythrocytes and their adhesion to endothelial receptors, obstructing blood flow to vital organs. Unlike other human-infecting Plasmodium species, P. falciparum exports a family of 18 FIKK serine/threonine kinases into the host cell, suggesting that phosphorylation may modulate erythrocyte modifications. We reveal substantial species-specific phosphorylation of erythrocyte proteins by P. falciparum but not by Plasmodium knowlesi, which does not export FIKK kinases. By conditionally deleting all FIKK kinases combined with large-scale quantitative phosphoproteomics we identified unique phosphorylation fingerprints for each kinase, including phosphosites on parasite virulence factors and host erythrocyte proteins. Despite their non-overlapping target sites, a network analysis revealed that some FIKKs may act in the same pathways. Only the deletion of the non-exported kinase FIKK8 resulted in reduced parasite growth, suggesting the exported FIKKs may instead support functions important for survival in the host. We show that one kinase, FIKK4.1, mediates both rigidification of the erythrocyte cytoskeleton and trafficking of the adhesin and key virulence factor PfEMP1 to the host cell surface. This establishes the FIKK family as important drivers of parasite evolution and malaria pathology.

Protein phosphorylation plays a key role in regulating biological processes. Over 30% of the proteome is phosphorylated in most organisms and unraveling the function of the kinases that mediate these phosphorylation events requires the technology to reliably measure phosphorylation on proteins under various conditions. Advances in mass-spectrometry instrumentation, sample preparation, and labeling technologies now offer a range of quantification methods, each with their advantages and disadvantages. Here we describe in detail two different quantification methods, that is, stable isotope labeling by amino acids in cell culture and tandem mass tagging, combined with phosphopeptide enrichment strategies to measure the phosphoproteome of Toxoplasma parasites.

Toxoplasma gondii contains a limited subset of actin binding proteins. Here we show that the putative actin regulator cyclase-associated protein (CAP) is present in two different isoforms and its deletion leads to significant defects in some but not all actin dependent processes. We observe defects in cell-cell communication, daughter cell orientation and the juxtanuclear accumulation of actin, but only modest defects in synchronicity of division and no defect in the replication of the apicoplast. 3D electron microscopy reveals that loss of CAP results in a defect in formation of a normal central residual body, but parasites remain connected within the vacuole. This dissociates synchronicity of division and parasite rosetting and reveals that establishment and maintenance of the residual body may be more complex than previously thought. These results highlight the different spatial requirements for F-actin regulation in Toxoplasma which appear to be achieved by partially overlapping functions of actin regulators., Competing Interests: AH, MR, JW, RK, RC, CP, LC, AH, GW, MT No competing interests declared, (© 2019, Hunt et al.)

Plasmodium falciparum has a complex life cycle that involves interaction with multiple tissues inside the human and mosquito hosts. Identification of essential genes at all different stages of the P. falciparum life cycle is urgently required for clinical development of tools for malaria control and eradication. However, the study of P. falciparum is limited by the inability to genetically modify the parasite throughout its life cycle with the currently available genetic tools. Here, we describe the detailed characterization of a new marker-free P. falciparum parasite line that expresses rapamycin-inducible Cre recombinase across the full life cycle. Using this parasite line, we were able to conditionally delete the essential invasion ligand AMA1 in three different developmental stages for the first time. We further confirm efficient gene deletion by targeting the nonessential kinase FIKK7.1. IMPORTANCE One of the major limitations in studying P. falciparum is that so far only asexual stages are amenable to rapid conditional genetic modification. The most promising drug targets and vaccine candidates, however, have been refractory to genetic modification because they are essential during the blood stage or for transmission in the mosquito vector. This leaves a major gap in our understanding of parasite proteins in most life cycle stages and hinders genetic validation of drug and vaccine targets. Here, we describe a method that supports conditional gene deletion across the P. falciparum life cycle for the first time. We demonstrate its potential by deleting essential and nonessential genes at different parasite stages, which opens up completely new avenues for the study of malaria and drug development. It may also allow the realization of novel vaccination strategies using attenuated parasites., (Copyright © 2019 Tibúrcio et al.)