Biology subjects

Mohsen, H.

Publications and source records attributed to Mohsen, H..

6 recordsLinked to original sources

bioRxiv2023

Dynamic clustering of genomics cohorts beyond race, ethnicity--and ancestry

BackgroundRecent decades have witnessed a steady decrease in the use of race categories in genomic studies. While studies that still include race categories vary in goal and type, these categories already build on a history during which racial color lines have been enforced and adjusted in the service of social and political systems of power and disenfranchisement. For early modern biological classification systems, data collection was also considerably arbitrary and limited. Fixed, discrete classifications have limited the study of human biodiversity and disrupted widely spread genetic and phenotypic continuums across geographic scales. Relatedly, the use of broad and predefined classification schemes--e.g. continent-based--across traits can risk missing important trait-specific genomic signals. ResultsTo address these issues, we introduce a dynamic approach to clustering human genomics cohorts on a trait-specific level and without using a set of predefined categories. We tested the approach on whole-exome sequencing datasets in ten cancer types and partitioned them based on germline variants in cancer-relevant genes that could confer cancer type-specific disease predisposition. Results demonstrate clustering patterns that transcend discrete continent-based categories across cancer types. Functional analysis based on cancer type-specific clusterings also captures the fundamental biological processes underlying cancer, differentiates between dynamic clusters on a functional level, and identifies novel potential drivers overlooked by a continent-based clustering model. ConclusionsThrough a trait-based lens, the dynamic clustering approach reveals genomic patterns that transcend predefined classification categories. We propose that coupled with diverse data collection, new clustering approaches have the potential to draw a more complete portrait of genomic variation and to address, in parallel, technical and social aspects of studying human biodiversity.

genomics↗

bioRxiv2021

A Network Propagation Approach to Prioritize Long Tail Genes in Cancer

IntroductionThe diversity of genomic alterations in cancer pose challenges to fully understanding the etiologies of the disease. Recent interest in infrequent mutations, in genes that reside in the "long tail" of the mutational distribution, uncovered new genes with significant implication in cancer development. The study of these genes often requires integrative approaches with multiple types of biological data. Network propagation methods have demonstrated high efficacy in uncovering genomic patterns underlying cancer using biological interaction networks. Yet, the majority of these analyses have focused their assessment on detecting known cancer genes or identifying altered subnetworks. In this paper, we introduce a network propagation approach that entirely focuses on long tail genes with potential functional impact on cancer development. ResultsWe identify sets of often overlooked, rarely to moderately mutated genes whose biological interactions significantly propel their mutation-frequency-based rank upwards during propagation in 17 cancer types. We call these sets "upward mobility genes" (UMGs, 28-83 genes per cancer type) and hypothesize that their significant rank improvement indicates functional importance. We report new cancer-pathway associations based on UMGs that were not previously identified using driver genes alone, validate UMGs role in cancer cell survival in vitro--alone and compared to other network methods--using extensive genome-wide RNAi and CRISPR data repositories, and further conduct in vitro functional screenings resulting the validation of 8 previously unreported genes. ConclusionOur analysis extends the spectrum of cancer relevant genes and identifies novel potential therapeutic targets.

bioinformatics↗

bioRxiv2021

Cancer relevance of human genes

BackgroundIt is unclear how many of genes contribute to the biology of cancer. We hypothesize that genes that interact with core cancer gene (CCG) in a protein-protein interaction network (PPI) may have functional importance. MethodsWe categorized genes into 1- (n=6791), 2- (n=7724), 3- (n=1587), and >3-steps (n=362) removed from the nearest CCG in the STRING PPI and demonstrate that the cancer-biology related functional contribution of the genes in these different neighborhood categories decreases as their distance from the CCGs increases. ResultsGenes closer to cancer genes manifest greater connectedness in the network, show greater importance in maintaining cell viability in a broad range of cancer cells in vitro, are also under greater negative germline selection pressure in the healthy populations, and have higher somatic mutation frequency and cancer effect. ConclusionsApproximately 70% of human genes are 1 or 2 steps removed from cancer genes in protein network and show functional importance in cancer-biology. These results suggest that the universe of cancer-relevant genes extends to thousands of genes that can contribute functional effects when dysregulated.

genomics↗

bioRxiv2020

Compression-based Network Interpretability Schemes

Deep learning methods have achieved state-of-the-art performance in many domains of artificial intelligence, but are typically hard to interpret. Network interpretation is important for multiple reasons, including knowledge discovery, hypothesis generation, fairness and establishing trust. Model transformations provide a general approach to interpreting a trained network post-hoc: the network is approximated by a model, which is typically compressed, whose structure can be more easily interpreted in some way (we call such approaches interpretability schemes). However, the relationship between compression and interpretation has not been fully explored: How much should a network be compressed for optimal extraction of interpretable information? Should compression be combined with other criteria when selecting model transformations? We investigate these issues using two different compression-based schemes, which aim to extract orthogonal kinds of information, pertaining to feature and data instance-based groupings respectively. The first (rank projection trees) uses a structured sparsification method such that nested groups of features can be extracted having potential joint interactions. The second (cascaded network decomposition) splits a network into a cascade of simpler networks, allowing groups of training instances with similar characteristics to be extracted at each stage of the cascade. We use predictive tasks in cancer and psychiatric genomics to assess the ability of these approaches to extract informative feature and data-point groupings from trained networks. We show that the generalization error of a network provides an indicator of the quality of the information extracted; further we derive PAC-Bayes generalization bounds for both schemes, which we show can be used as proxy indicators, and can thus provide a criterion for selecting the optimal compression. Finally, we show that the PAC-Bayes framework can be naturally modified to incorporate additional criteria alongside compression, such as prior knowledge based on previous models, which can enhance interpretable model selection.

bioinformatics↗

bioRxiv2018

GRAM: A generalized model to predict the molecular effect of a non-coding variant in a cell-type specific manner

There has been much effort to prioritize genomic variants with respect to their impact on \"function\". However, function is often not precisely defined: Sometimes, it is the disease association of a variant; other times, it reflects a molecular effect on transcription or epigenetics. Here we coupled multiple genomic predictors to build GRAM, a generalized model, to predict a well-defined experimental target: the expression-modulating effect of a non-coding variant in a cell-specific manner. As a first step, we performed feature engineering: using a LASSO regularized linear model, we found transcription factor (TF) binding most predictive, especially for TFs that are hubs in the regulatory network; in contrast, evolutionary conservation, a popular feature in many other functional-impact predictors, has almost no contribution. Moreover, TF binding inferred from in vitro SELEX is as effective as that from in vivo ChIP-Seq. Second, we implemented GRAM integrating SELEX features and expression profiles. The program combines a universal regulatory score for a variant in a non-coding element with a modifier score reflecting the particular cell type. We benchmarked GRAM on a large-scale MPRA dataset in the GM12878 cell line, achieving a ROC score of [~]0.73; performance on the K562 cell line was similar. Finally, we evaluated the performance of GRAM on targeted regions using luciferase assays in MCF7 and K562 cell lines. We noted that changing the insertion position of the construct relative to the reporter gene gives very different results, highlighting the importance of carefully defining the functional target the model is predicting.\n\nAuthor SummaryNoncoding variants lie outside of protein-coding regions, and are found to have disease associations. However, knowledge on the molecular effect of these non-coding variants in a cell-specific context is very limited. Also, different output between multiple experiment platforms may introduce extra complexity in analyzing the molecular function of these variants. We developed GRAM, a generalized model to predict molecular effect of non-coding variants in multiple cell types for different experimental platforms. We first selected the most informative cell-independent SELEX transcription factor binding score on the variant locus as features and then combine cell-specific gene expression profile to build a multi-step prediction model. GRAM has been successfully tested on both MPRA and Luciferase assay, and on three different cell lines: GM12878, K562 and MCF7, shows high performance.

bioinformatics↗

bioRxiv2017

Exploring Single-Cell Data with Multitasking Deep Neural Networks

Biomedical researchers are generating high-throughput, high-dimensional single-cell data at a staggering rate. As costs of data generation decrease, experimental design is moving towards measurement of many different single-cell samples in the same dataset. These samples can correspond to different patients, conditions, or treatments. While scalability of methods to datasets of these sizes is a challenge on its own, dealing with large-scale experimental design presents a whole new set of problems, including batch effects and sample comparison issues. Currently, there are no computational tools that can both handle large amounts of data in a scalable manner (many cells) and at the same time deal with many samples (many patients or conditions). Moreover, data analysis currently involves the use of different tools that each operate on their own data representation, not guaranteeing a synchronized analysis pipeline. For instance, data visualization methods can be disjoint and mismatched with the clustering method. For this purpose, we present SAUCIE, a deep neural network that leverages the high degree of parallelization and scalability offered by neural networks, as well as the deep representation of data that can be learned by them to perform many single-cell data analysis tasks, all on a unified representation.\n\nA well-known limitation of neural networks is their interpretability. Our key contribution here are newly formulated regularizations (penalties) that render features learned in hidden layers of the neural network interpretable. When large multi-patient datasets are fed into SAUCIE, the various hidden layers contain denoised and batch-corrected data, a low dimensional visualization, unsupervised clustering, as well as other information that can be used to explore the data. We show this capability by analyzing a newly generated 180-sample dataset consisting of T cells from dengue patients in India, measured with mass cytometry. We show that SAUCIE, for the first time, can batch correct and process this 11-million cell data to identify cluster-based signatures of acute dengue infection and create a patient manifold, stratifying immune response to dengue on the basis of single-cell measurements.

bioinformatics↗