Computational Construction of Toxicant Signaling Networks.

Humans and animals are regularly exposed to compounds that may have adverse effects on health. The Toxicity Forecaster (ToxCast) program was developed to use high throughput screening assays to quickly screen chemicals by measuring their effects on many biological end points. Many of these assays test for effects on cellular receptors and transcription factors (TFs), under the assumption that a toxicant may perturb normal signaling pathways in the cell. We hypothesized that we could reconstruct the intermediate proteins in these pathways that may be directly or indirectly affected by the toxicant, potentially revealing important physiological processes not yet tested for many chemicals. We integrate data from ToxCast with a human protein interactome to build toxicant signaling networks that contain physical and signaling protein interactions that may be affected as a result of toxicant exposure. To build these networks, we developed the EdgeLinker algorithm, which efficiently finds short paths in the interactome that connect the receptors to TFs for each toxicant. We performed multiple evaluations and found evidence suggesting that these signaling networks capture biologically relevant effects of toxicants. To aid in dissemination and interpretation, interactive visualizations of these networks are available at http://graphspace.org.

Upper-Bound Energy Minimization to Search for Stable Functional Materials with Graph Neural Networks.

The discovery of new materials in unexplored chemical spaces necessitates quick and accurate prediction of thermodynamic stability, often assessed using density functional theory (DFT), and efficient search strategies. Here, we develop a new approach to finding stable inorganic functional materials. We start by defining an upper bound to the fully relaxed energy obtained via DFT as the energy resulting from a constrained optimization over only cell volume. Because the fractional atomic coordinates for these calculations are known a priori, this upper bound energy can be quickly and accurately predicted with a scale-invariant graph neural network (GNN). We generate new structures via ionic substitution of known prototypes, and train our GNN on a new database of 128 000 DFT calculations comprising both fully relaxed and volume-only relaxed structures. By minimizing the predicted upper-bound energy, we discover new stable structures with over 99% accuracy (versus DFT). We demonstrate the method by finding promising new candidates for solid-state battery (SSB) electrolytes that not only possess the required stability, but also additional functional properties such as large electrochemical stability windows and high conduction ion fraction. We expect this proposed framework to be directly applicable to a wide range of design challenges in materials science.

Integrating multimodal data through interpretable heterogeneous ensembles.

Motivation: Integrating multimodal data represents an effective approach to predicting biomedical characteristics, such as protein functions and disease outcomes. However, existing data integration approaches do not sufficiently address the heterogeneous semantics of multimodal data. In particular, early and intermediate approaches that rely on a uniform integrated representation reinforce the consensus among the modalities but may lose exclusive local information. The alternative late integration approach that can address this challenge has not been systematically studied for biomedical problems. Results: We propose Ensemble Integration (EI) as a novel systematic implementation of the late integration approach. EI infers local predictive models from the individual data modalities using appropriate algorithms and uses heterogeneous ensemble algorithms to integrate these local models into a global predictive model. We also propose a novel interpretation method for EI models. We tested EI on the problems of predicting protein function from multimodal STRING data and mortality due to coronavirus disease 2019 (COVID-19) from multimodal data in electronic health records. We found that EI accomplished its goal of producing significantly more accurate predictions than each individual modality. It also performed better than several established early integration methods for each of these problems. The interpretation of a representative EI model for COVID-19 mortality prediction identified several disease-relevant features, such as laboratory test (blood urea nitrogen and calcium) and vital sign measurements (minimum oxygen saturation) and demographics (age). These results demonstrated the effectiveness of the EI framework for biomedical data integration and predictive modeling. Availability and implementation: Code and data are available at https://github.com/GauravPandeyLab/ensemble_integration. Supplementary information: Supplementary data are available at Bioinformatics Advances online.

Integrating multimodal data through interpretable heterogeneous ensembles.

Motivation: Integrating multimodal data represents an effective approach to predicting biomedical characteristics, such as protein functions and disease outcomes. However, existing data integration approaches do not sufficiently address the heterogeneous semantics of multimodal data. In particular, early and intermediate approaches that rely on a uniform integrated representation reinforce the consensus among the modalities, but may lose exclusive local information. The alternative late integration approach that can address this challenge has not been systematically studied for biomedical problems. Results: We propose Ensemble Integration (EI) as a novel systematic implementation of the late integration approach. EI infers local predictive models from the individual data modalities using appropriate algorithms, and uses effective heterogeneous ensemble algorithms to integrate these local models into a global predictive model. We also propose a novel interpretation method for EI models. We tested EI on the problems of predicting protein function from multimodal STRING data, and mortality due to COVID-19 from multimodal data in electronic health records. We found that EI accomplished its goal of producing significantly more accurate predictions than each individual modality. It also performed better than several established early integration methods for each of these problems. The interpretation of a representative EI model for COVID-19 mortality prediction identified several disease-relevant features, such as laboratory test (blood urea nitrogen (BUN) and calcium) and vital sign measurements (minimum oxygen saturation) and demographics (age). These results demonstrated the effectiveness of the EI framework for biomedical data integration and predictive modeling. Availability: Code and data are available at https://github.com/GauravPandeyLab/ensemble_integration . Contact: gaurav.pandey@mssm.edu.

Interpretable network propagation with application to expanding the repertoire of human proteins that interact with SARS-CoV-2.

BACKGROUND: Network propagation has been widely used for nearly 20 years to predict gene functions and phenotypes. Despite the popularity of this approach, little attention has been paid to the question of provenance tracing in this context, e.g., determining how much any experimental observation in the input contributes to the score of every prediction. RESULTS: We design a network propagation framework with 2 novel components and apply it to predict human proteins that directly or indirectly interact with SARS-CoV-2 proteins. First, we trace the provenance of each prediction to its experimentally validated sources, which in our case are human proteins experimentally determined to interact with viral proteins. Second, we design a technique that helps to reduce the manual adjustment of parameters by users. We find that for every top-ranking prediction, the highest contribution to its score arises from a direct neighbor in a human protein-protein interaction network. We further analyze these results to develop functional insights on SARS-CoV-2 that expand on known biology such as the connection between endoplasmic reticulum stress, HSPA5, and anti-clotting agents. CONCLUSIONS: We examine how our provenance-tracing method can be generalized to a broad class of network-based algorithms. We provide a useful resource for the SARS-CoV-2 community that implicates many previously undocumented proteins with putative functional relationships to viral infection. This resource includes potential drugs that can be opportunistically repositioned to target these proteins. We also discuss how our overall framework can be extended to other, newly emerging viruses.

Protein sequence models for prediction and comparative analysis of the SARS-CoV-2 -human interactome.

Viruses such as the novel coronavirus, SARS-CoV-2, that is wreaking havoc on the world, depend on interactions of its own proteins with those of the human host cells. Relatively small changes in sequence such as between SARS-CoV and SARS-CoV-2 can dramatically change clinical phenotypes of the virus, including transmission rates and severity of the disease. On the other hand, highly dissimilar virus families such as Coronaviridae, Ebola, and HIV have overlap in functions. In this work we aim to analyze the role of protein sequence in the binding of SARS-CoV-2 virus proteins towards human proteins and compare it to that of the above other viruses. We build supervised machine learning models, using Generalized Additive Models to predict interactions based on sequence features and find that our models perform well with an AUC-PR of 0.65 in a class-skew of 1:10. Analysis of the novel predictions using an independent dataset showed statistically significant enrichment. We further map the importance of specific amino-acid sequence features in predicting binding and summarize what combinations of sequences from the virus and the host is correlated with an interaction. By analyzing the sequence-based embeddings of the interactomes from different viruses and clustering them together we find some functionally similar proteins from different viruses. For example, vif protein from HIV-1, vp24 from Ebola and orf3b from SARS-CoV all function as interferon antagonists. Furthermore, we can differentiate the functions of similar viruses, for example orf3a's interactions are more diverged than orf7b interactions when comparing SARS-CoV and SARS-CoV-2.

Development, implementation, and evaluation of a competency-based didactic and simulation-focused boot camp for incoming urology residents: Report of the first three years.

INTRODUCTION: The Royal College of Physicians and Surgeons of Canada's Competence by Design (CBD) initiative presents curricula challenges to ensure residents gain proficiency while progressing through training. To prepare first-year urology residents (R1s), we developed, implemented, and evaluated a didactic and simulation-focused boot camp to implement the CBD curriculum. We report our experiences and findings of the first three years. METHODS: Urology residents from two Canadian universities participated in the two-day boot camp at the beginning of residency. Eleven didactic and six simulation sessions allowed for instruction and deliberate practice with feedback. Pre-and post-course multiple-choice questionnaires (MCQs) and an objective structured clinical exam (OSCE) evaluated knowledge and skills uptake. For initial program evaluation, three R2s served as historical controls in year 1. RESULTS: Nineteen residents completed boot camp. The mean age was 26.4 (±2.8) and 13 were male. Participants markedly improved on the pre- and post-MCQs (year 1: 62% and 91%; year 2: 55% and 89%; year 3: 58% and 86%, respectively). Participants scored marginally higher than the controls on four of the six OSCE stations. OSCE scores remained >88% over the three cohorts. All participants reported higher confidence levels post-boot camp and felt it was excellent preparation for residency. CONCLUSIONS: During its first three years, our urology boot camp has demonstrated high feasibility and utility. Knowledge and technical skills uptake were established via MCQ and OSCE results, with participants' scores near or above those of R2 controls. This boot camp will remain in our CBD curriculum and can provide a framework for other urology residency programs.

Accurate and efficient gene function prediction using a multi-bacterial network.

MOTIVATION: Nearly 40% of the genes in sequenced genomes have no experimentally or computationally derived functional annotations. To fill this gap, we seek to develop methods for network-based gene function prediction that can integrate heterogeneous data for multiple species with experimentally based functional annotations and systematically transfer them to newly sequenced organisms on a genome-wide scale. However, the large sizes of such networks pose a challenge for the scalability of current methods. RESULTS: We develop a label propagation algorithm called FastSinkSource. By formally bounding its rate of progress, we decrease the running time by a factor of 100 without sacrificing accuracy. We systematically evaluate many approaches to construct multi-species bacterial networks and apply FastSinkSource and other state-of-the-art methods to these networks. We find that the most accurate and efficient approach is to pre-compute annotation scores for species with experimental annotations, and then to transfer them to other organisms. In this manner, FastSinkSource runs in under 3 min for 200 bacterial species. AVAILABILITY AND IMPLEMENTATION: An implementation of our framework and all data used in this research are available at https://github.com/Murali-group/multi-species-GOA-prediction. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Benchmarking algorithms for gene regulatory network inference from single-cell transcriptomic data.

We present a systematic evaluation of state-of-the-art algorithms for inferring gene regulatory networks from single-cell transcriptional data. As the ground truth for assessing accuracy, we use synthetic networks with predictable trajectories, literature-curated Boolean models and diverse transcriptional regulatory networks. We develop a strategy to simulate single-cell transcriptional data from synthetic and Boolean networks that avoids pitfalls of previously used methods. Furthermore, we collect networks from multiple experimental single-cell RNA-seq datasets. We develop an evaluation framework called BEELINE. We find that the area under the precision-recall curve and early precision of the algorithms are moderate. The methods are better in recovering interactions in synthetic networks than Boolean models. The algorithms with the best early precision values for Boolean models also perform well on experimental datasets. Techniques that do not require pseudotime-ordered cells are generally more accurate. Based on these results, we present recommendations to end users. BEELINE will aid the development of gene regulatory network inference algorithms.

Integrating protein localization with automated signaling pathway reconstruction.

BACKGROUND: Understanding cellular responses via signal transduction is a core focus in systems biology. Tools to automatically reconstruct signaling pathways from protein-protein interactions (PPIs) can help biologists generate testable hypotheses about signaling. However, automatic reconstruction of signaling pathways suffers from many interactions with the same confidence score leading to many equally good candidates. Further, some reconstructions are biologically misleading due to ignoring protein localization information. RESULTS: We propose LocPL, a method to improve the automatic reconstruction of signaling pathways from PPIs by incorporating information about protein localization in the reconstructions. The method relies on a dynamic program to ensure that the proteins in a reconstruction are localized in cellular compartments that are consistent with signal transduction from the membrane to the nucleus. LocPL and existing reconstruction algorithms are applied to two PPI networks and assessed using both global and local definitions of accuracy. LocPL produces more accurate and biologically meaningful reconstructions on a versatile set of signaling pathways. CONCLUSION: LocPL is a powerful tool to automatically reconstruct signaling pathways from PPIs that leverages cellular localization information about proteins. The underlying dynamic program and signaling model are flexible enough to study cellular signaling under different settings of signaling flow across the cellular compartments.

Missed opportunities for DCD kidney donors: Evaluation of warm ischemic time and associated functional warm ischemic time.

INTRODUCTION: Many transplant centers utilize a hard cutoff of 2 hours of warm ischemic time (WIT), defined as the time from withdrawal of life-sustaining measures to cold organ flush, to exclude donation after circulatory determination of death (DCD) kidney donation. As a result, almost a quarter of withdrawals to retrieve DCD organs fail to produce kidney transplants in Ontario. In order to assess our ability to increase organ yield, we wanted to characterize WIT and functional WIT (fWIT, time from systolic blood pressure <50 mm Hg to cold organ flush), as well as determine the time at which potential donors eventually die in those that did not become organ donors. METHODS: A retrospective review of all DCD kidney donors in Ontario was performed utilizing the Trillium Gift of Life Database from April 2013 to February 2018. RESULTS: Of 350 DCD kidney donors analyzed, 46.9% had < 0.5 hours, 51.7% between 0.5 and 2 hours, and 1.4% >2 hours of WIT. In each of these categories (WIT <0.5 hours, 0.5-2 hours and >2 hours), the percentage of patients with fWIT <30 minutes was 100%, 94.4%, and 100%, respectively (P = NS). There were 106 potential donors who did not end up donating due to WIT >2 hours. Of these, 20.8% died between 2 and 4 hours, 10.4% between 4 and 6 hours, and 68.8% beyond 6 hours. DISCUSSION: The percentage of donors with fWIT >30 minutes did not increase with increasing WIT in DCD donors that went on to donate organs. These data support assessment of waiting up to 4 hours for DCD kidney donation as long as fWIT remains low.

Significance of atypical urinary cytology in the evaluation of patients with end-stage renal disease for kidney transplantation - a retrospective study.

To determine what percentage of renal transplant candidates have atypical urinary cytology, what proportion have urothelial carcinoma and whether cystoscopy is necessary with atypical cytology. All end-stage renal disease (ESRD) patients (703) presenting for renal transplantation at our institution were retrospectively reviewed. Individuals producing sufficient urine were screened with urine cytology and those with atypical cytology or risk factors for bladder cancer underwent cystoscopy. Four hundred and thirty patients had available urinary cytology and, of these, 151 (35%) had atypical cytology. Of patients with atypical cytology, three were identified to have urothelial carcinoma. However, three additional patients with urothelial carcinoma did not present with atypical cytology. In total, 6 of 703 (0.85%) patients had bladder cancer. All were treated with transurethral resection and eventually underwent renal transplant. One patient has had disease progression post-transplant to distant metastases. This is the largest study to date evaluating the incidence of urothelial carcinoma in ESRD patients presenting for transplant workup. We found the incidence of bladder cancer to be higher than in the general Canadian population, however, most lesions were low grade. We found atypical cytology in transplant candidates to be a poor predictor for these low-grade lesions and do not recommend routine cystoscopy for atypical cytology.

Large-scale protein function prediction using heterogeneous ensembles.

Heterogeneous ensembles are an effective approach in scenarios where the ideal data type and/or individual predictor are unclear for a given problem. These ensembles have shown promise for protein function prediction (PFP), but their ability to improve PFP at a large scale is unclear. The overall goal of this study is to critically assess this ability of a variety of heterogeneous ensemble methods across a multitude of functional terms, proteins and organisms. Our results show that these methods, especially Stacking using Logistic Regression, indeed produce more accurate predictions for a variety of Gene Ontology terms differing in size and specificity. To enable the application of these methods to other related problems, we have publicly shared the HPC-enabled code underlying this work as LargeGOPred ( https://github.com/GauravPandeyLab/LargeGOPred).

Automating the PathLinker app for Cytoscape.

PathLinker is a graph-theoretic algorithm originally developed to reconstruct the interactions in a signaling pathway of interest. It efficiently computes multiple short paths within a background protein interaction network from the receptors to transcription factors (TFs) in a pathway. Since December 2015, PathLinker has been available as an app for Cytoscape. This paper describes how we automated the app to use the CyRest infrastructure and how users can incorporate PathLinker into their software pipelines.

The PathLinker app: Connect the dots in protein interaction networks.

PathLinker is a graph-theoretic algorithm for reconstructing the interactions in a signaling pathway of interest. It efficiently computes multiple short paths within a background protein interaction network from the receptors to transcription factors (TFs) in a pathway. We originally developed PathLinker to complement manual curation of signaling pathways, which is slow and painstaking. The method can be used in general to connect any set of sources to any set of targets in an interaction network. The app presented here makes the PathLinker functionality available to Cytoscape users. We present an example where we used PathLinker to compute and analyze the network of interactions connecting proteins that are perturbed by the drug lovastatin.

First Canadian experience with robotic single-incision pyeloplasty: Comparison with multi-incision technique.

INTRODUCTION: We compared the outcomes of single-incision, robot-assisted laparoscopic pyeloplasty vs. multiple-incision pyeloplasty using the da Vinci robotic system. METHODS: We reviewed all consecutive robotic pyeloplasties by a single surgeon from January 2011 to August 2015. A total of 30 procedures were performed (16 single:14 multi-port). Two different single-port devices were compared: the GelPort (Applied Medical, Rancho Santa Margarita, CA) and the Intuitive single-site access port (Intuitive Surgical, Sunnyvale, CA). RESULTS: Patient demographics were similar between the two groups. Mean operating time was similar among the single and multi-port groups (225.2 min vs. 198.9 minutes [p=0.33]). There was no significant difference in length of hospital stay in either group (86.2 hr vs. 93.2 hr [p=0.76]). There was no difference in success rates or postoperative complications among groups. CONCLUSIONS: Single-port robotic pyeloplasty is non-inferior to multiple-incision robotic surgery in terms of operative times, hospitalization time, success rates, and complications. Verifying these results with larger cohorts is required prior to the wide adoption of this technique. Ongoing objective measurements of cosmesis and patient satisfaction are being evaluated.

Nodal signals via ß-arrestins and RalGTPases to regulate trophoblast invasion.

Placentation is critical for establishing a healthy pregnancy. Trophoblasts mediate implantation and placentation and certain subtypes, most notably extravillous cytotrophoblast, are highly invasive. Trophoblast invasion is tightly regulated by microenvironmental cues that dictate placental morphology and depth. In choriocarcinomas, malignant trophoblast cells become hyperinvasive, breaching the myometrium and leading to major complications. Nodal, a member of the TGF-ß superfamily, is expressed throughout the endometrium during the peri-implantation period and in invasive trophoblast cells. Nodal promotes the invasion of numerous types of cancer cells. However, Nodal's role in trophoblast and choriocarcinoma cell invasion is unclear. Here we show that Nodal stimulates the invasion of both the non-malignant HTR-8SV/neo trophoblast and JAR choriocarcinoma cells in a dose-dependent manner. We found that endogenous ß-arrestins and Ral GTPases, key regulators of the cell cytoskeleton, are constitutively associated with Nodal receptors (ALK4 and ALK7) in trophoblast cells and that RalA is colocalized with ALK4 in endocytic vesicles. Nodal stimulates endogenous ß-arrestin2 to associate with phospho-ERK1/2, and knockdown of ß-arrestin or Ral proteins impairs Nodal-induced trophoblast and choriocarcinoma cell invasion. These results demonstrate, for the first time, that ß-arrestins and RalGTPases are important regulators of Nodal-induced invasion.

GPR54 (KISS1R) transactivates EGFR to promote breast cancer cell invasiveness.

Kisspeptins (Kp), peptide products of the Kisspeptin-1 (KISS1) gene are endogenous ligands for a G protein-coupled receptor 54 (GPR54). Previous findings have shown that KISS1 acts as a metastasis suppressor in numerous cancers in humans. However, recent studies have demonstrated that an increase in KISS1 and GPR54 expression in human breast tumors correlates with higher tumor grade and metastatic potential. At present, whether or not Kp signaling promotes breast cancer cell invasiveness, required for metastasis and the underlying mechanisms, is unknown. We have found that kisspeptin-10 (Kp-10), the most potent Kp, stimulates the invasion of human breast cancer MDA-MB-231 and Hs578T cells using Matrigel-coated Transwell chamber assays and induces the formation of invasive stellate structures in three-dimensional invasion assays. Furthermore, Kp-10 stimulated an increase in matrix metalloprotease (MMP)-9 activity. We also found that Kp-10 induced the transactivation of epidermal growth factor receptor (EGFR). Knockdown of the GPCR scaffolding protein, ß-arrestin 2, inhibited Kp-10-induced EGFR transactivation as well as Kp-10 induced invasion of breast cancer cells via modulation of MMP-9 secretion and activity. Finally, we found that the two receptors associate with each other under basal conditions, and FRET analysis revealed that GPR54 interacts directly with EGFR. The stability of the receptor complex formation was increased upon treatment of cells by Kp-10. Taken together, our findings suggest a novel mechanism by which Kp signaling via GPR54 stimulates breast cancer cell invasiveness.

The effectiveness of training with an emergency department simulator on medical student performance in a simulated disaster.

OBJECTIVE: Training in practical aspects of disaster medicine is often impossible, and simulation may offer an educational opportunity superior to traditional didactic methods. We sought to determine whether exposure to an electronic simulation tool would improve the ability of medical students to manage a simulated disaster. METHODS: We stratified 22 students by year of education and randomly assigned 50% from each category to form the intervention group, with the remaining 50% forming the control group. Both groups received the same didactic training sessions. The intervention group received additional disaster medicine training on a patient simulator (disastermed.ca), and the control group spent equal time on the simulator in a nondisaster setting. We compared markers of patient flow during a simulated disaster, including mean differences in time and number of patients to reach triage, bed assignment, patient assessment and disposition. In addition, we compared triage accuracy and scores on a structured command-and-control instrument. We collected data on the students' evaluations of the course for secondary purposes. RESULTS: Participants in the intervention group triaged their patients more quickly than participants in the control group (mean difference 43 s, 99.5% confidence interval [CI] 12 to 75 s). The score of performance indicators on a standardized scale was also significantly higher in the intervention group (18/18) when compared with the control group (8/18) (p < 0.001). All students indicated that they preferred the simulation-based curriculum to a lecture-based curriculum. When asked to rate the exercise overall, both groups gave a median score of 8 on a 10-point modified Likert scale. CONCLUSION: Participation in an electronic disaster simulation using the disastermed.ca software package appears to increase the speed at which medical students triage simulated patients and increase their score on a structured command-and-control performance indicator instrument. Participants indicated that the simulation-based curriculum in disaster medicine is preferable to a lecture-based curriculum. Overall student satisfaction with the simulation-based curriculum was high.