Biology subjects

Ravanmehr, V.

Publications and source records attributed to Ravanmehr, V..

8 recordsLinked to original sources

bioRxiv2024

Diverse ancestral representation improves genetic intolerance metrics

Rapidly expanding genomic databases have enabled the identification of regions in the human genome intolerant to variation and thus likely relevant to human disease. However, despite their unprecedented scale, these datasets remain constrained by limited ancestral diversity. Here, we systematically evaluate how genetic diversity impacts gene- and sub-genic intolerance metrics by analyzing whole-exome sequencing data from 460,551 UK Biobank participants and 125,748 gnomAD participants across diverse ancestral backgrounds. Through comprehensive analysis of randomly sampled datasets with varying ancestral compositions, we demonstrate that genetic diversity, rather than sample size alone, drives the performance of intolerance metrics in identifying functionally critical genes and genic sub-regions. Notably, scores trained on variation observed in African and Admixed American ancestral groups showed superior resolution in detecting neurodevelopmental disease risk genes and haploinsufficient genes compared to scores trained on variation observed in European ancestry cohorts. Most strikingly, the Missense Tolerance Ratio (MTR) trained on 43,000 multi-ancestry exomes demonstrated greater predictive power than when trained on a nearly 10-fold larger dataset of 440,000 non-Finnish European exomes, indicating that European ancestry-based scores are approaching saturation. These findings establish ancestral diversity as the fundamental determinant for advancing intolerance metrics and underscore the urgent scientific and ethical imperative for enhanced population representation in genomic resources to fully realize the potential of precision medicine and drug discovery.

genetics↗

bioRxiv2023

CREBBP lysine acetyltransferase domain mutations create zombie enzymes that alter chromatin loading dynamics and prevent EP300 redundancy

Epigenetic regulation of gene expression is fundamental to cell state transitions. A prominent example of this is the epigenetically-regulated generation of transitory cells states within the germinal center (GC) reaction that are required for humoral immunity. The deregulation of these processes through somatic mutations can give rise to B-cell lymphoma, thus studying these mutations can provide insight into fundamental mechanisms underlying epigenetic regulation of gene expression as well as GC B-cell and lymphoma biology. Here we show that classes of mutations in the CREBBP lysine acetyltransferase gene result in different structure, catalytic activity and function. We discover that CREBBP and its paralog EP300 are dynamically and reciprocally loaded onto chromatin during a signal-responsive cell state transitions within the GC. Mutation of CREBBP in GC-derived lymphoma interrupt signal-responsive CREBBP loading onto chromatin, while also inhibiting EP300 redundancy. These observations lead to a model in which CREBBP or EP300 loss of function through acetyltransferase domain mutations, which occur frequently in both hematopoietic and solid tumors, create a zombie enzyme that retards the function of the active paralog by binding limiting concentrations of transcription factor substrate and preventing signal-responsive enhancer activation.

cancer biology↗

bioRxiv2022

Node-degree aware edge sampling mitigates inflated classification performance in biomedical graph representation learning

Graph representation learning is a family of related approaches that learn low-dimensional vector representations of nodes and other graph elements called embeddings. Embeddings approximate characteristics of the graph and can be used for a variety of machine-learning tasks such as novel edge prediction. For many biomedical applications, partial knowledge exists about positive edges that represent relationships between pairs of entities, but little to no knowledge is available about negative edges that represent the explicit lack of a relationship between two nodes. For this reason, classification procedures are forced to assume that the vast majority of unlabeled edges are negative. Existing approaches to sampling negative edges for training and evaluating classifiers do so by uniformly sampling pairs of nodes. We show here that this sampling strategy typically leads to sets of positive and negative edges with imbalanced edge degree distributions. Using representative homogeneous and heterogeneous biomedical knowledge graphs, we show that this strategy artificially inflates measured classification performance. We present a degree-aware node sampling approach for sampling negative edge examples that mitigates this effect and is simple to implement.

bioinformatics↗

bioRxiv2021

Supervised learning with word embeddings derived from PubMed captures latent knowledge about protein kinases and cancer

Inhibiting protein kinases (PKs) that cause cancers has been an important topic in cancer therapy for years. So far, almost 8% of more than 530 PKs have been targeted by FDA-approved medications and around 150 protein kinase inhibitors (PKIs) have been tested in clinical trials. We present an approach based on natural language processing and machine learning to the relations between PKs and cancers, predicting PKs whose inhibition would be efficacious to treat a certain cancer. Our approach represents PKs and cancers as semantically meaningful 100-dimensional vectors based on co-occurrence patterns in PubMed abstracts. We use information about phase I-IV trials in ClinicalTrials.gov to construct a training set for random forest classification. In historical data, associations between PKs and specific cancers could be predicted years in advance with good accuracy. Our model may be a tool to predict the relevance of inhibiting PKs with specific cancers.

bioinformatics↗

bioRxiv2020

KG-COVID-19: a framework to produce customized knowledge graphs for COVID-19 response

Integrated, up-to-date data about SARS-CoV-2 and coronavirus disease 2019 (COVID-19) is crucial for the ongoing response to the COVID-19 pandemic by the biomedical research community. While rich biological knowledge exists for SARS-CoV-2 and related viruses (SARS-CoV, MERS-CoV), integrating this knowledge is difficult and time consuming, since much of it is in siloed databases or in textual format. Furthermore, the data required by the research community varies drastically for different tasks - the optimal data for a machine learning task, for example, is much different from the data used to populate a browsable user interface for clinicians. To address these challenges, we created KG-COVID-19, a flexible framework that ingests and integrates biomedical data to produce knowledge graphs (KGs) for COVID-19 response. This KG framework can also be applied to other problems in which siloed biomedical data must be quickly integrated for different research applications, including future pandemics. BIGGER PICTUREAn effective response to the COVID-19 pandemic relies on integration of many different types of data available about SARS-CoV-2 and related viruses. KG-COVID-19 is a framework for producing knowledge graphs that can be customized for downstream applications including machine learning tasks, hypothesis-based querying, and browsable user interface to enable researchers to explore COVID-19 data and discover relationships.

bioinformatics↗

bioRxiv2019

Perturbation of effector and regulatory T cell subsets in Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS)

Myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) is a debilitating disorder of unknown etiology, and diagnosis of the disease is largely based on clinical symptoms. We hypothesized that immunological disruption is the major driver of this disease and analyzed a large cohort of ME/CFS patient or control blood samples for differences in T cell subset frequencies and functions. We found that the ratio of CD4+ to CD8+ T cells and the proportion of CD8+ effector memory T cells were increased, whereas NK cells were reduced in ME/CFS patients younger than 50 years old compared to a healthy control group. Remarkably, major differences were observed in Th1, Th2, Th17 and mucosal-associated invariant T (MAIT) T cell subset functions across all ages of patients compared to healthy subjects. While CCR6+ Th17 cells in ME/CFS secreted less IL-17 compared to controls, their overall frequency was higher. Similarly, MAIT cells from patients secreted lower IFN{gamma}, GranzymeA and IL-17 upon activation. Together, these findings suggest chronic stimulation of these T cell populations in ME/CFS patients. In contrast, the frequency of regulatory T cells (Tregs), which control excessive immune activation, was higher in ME/CFS patients. Finally, using a machine learning algorithm called random forest, we determined that the set of T cell parameters analyzed could identify more than 90% of the subjects in the ME/CFS cohort as patients (93% true positive rate or sensitivity). In conclusion, these multiple and major perturbations or dysfunctions in T cell subsets in ME/CFS patients suggest potential chronic infections or microbiome dysbiosis. These findings also have implications for development of ME/CFS specific immune biomarkers and reveal potential targets for novel therapeutic interventions.

immunology↗

bioRxiv2019

Semantic Integration of Clinical Laboratory Tests from Electronic Health Records for Deep Phenotyping and Biomarker Discovery

Electronic Health Record (EHR) systems typically define laboratory test results using the Laboratory Observation Identifier Names and Codes (LOINC) and can transmit them using Fast Healthcare Interoperability Resource (FHIR) standards. LOINC has not yet been semantically integrated with computational resources for phenotype analysis. Here, we provide a method for mapping LOINC-encoded laboratory test results transmitted in FHIR standards to the Human Phenotype Ontology (HPO) terms. We annotated the medical implications of 2421 commonly used laboratory tests with HPO terms. Using these annotations, a software assesses laboratory test results and converts each into an HPO term. We validated our approach with EHR data from 15,681 patients with respiratory complaints and identified known biomarkers for asthma. Finally, we provide a freely available SMART on FHIR application that can be used within EHR systems. Our approach allows reusing readily available laboratory tests in EHR for deep phenotyping and using the hierarchical structure of HPO for association studies with medical outcomes and genomics.\n\nOne Sentence SummaryWe present an approach to semantically integrating LOINC-encoded laboratory data with the Human Phenotype Ontology and show that the integrated LOINC data can be used to identify biomarkers for asthma from electronic health record data.

bioinformatics↗

bioRxiv2017

ChIPWig: A Random Access-Enabling Lossless And Lossy Compression Method For ChIP-Seq Data

MotivationThe past decade has witnessed a rapid development of data acquisition technologies that enable integrative genomic and proteomic analysis. One such technology is chromatin immunoprecipitation sequencing (ChIP-seq), developed for analyzing interactions between proteins and DNA via next-generation sequencing technologies. As ChIP-seq experiments are inexpensive and time-efficient, massive datasets from this domain have been acquired, introducing significant storage and maintenance challenges. To address the resulting Big Data problems, we propose a state-of-the-art lossless and lossy compression framework specifically designed for ChIP-seq Wig data, termed ChIPWig. Wig is a standard file format, which in this setting contains relevant read density information crucial for visualization and downstream processing. ChIPWig may be executed in two different modes: lossless and lossy. Lossless ChIPWig compression allows for random access and fast queries in the file through careful variable-length block-wise encoding. ChIPWig also stores the summary statistics of each block needed for guided access. Lossy ChIPWig, in contrast, performs quantization of the read density values before feeding them into the lossless ChIPWig compressor. Nonuniform lossy quantization leads to further reductions in the file size, while maintaining the same accuracy of the ChIP-seq peak calling and motif discovery pipeline based on the NarrowPeaks method tailor-made for Wig files. The compressors are designed using new statistical modeling approaches coupled with delta and arithmetic encoding.\n\nResultsWe tested the ChIPWig compressor on a number of ChIP-seq datasets generated by the ENCODE project. Lossless ChIPWig reduces the file sizes to merely 6% of the original, and offers an average 6-fold compression rate improvement compared to bigWig. The running times for compression and decompression are comparable to those of bigWig. The compression and decompression speed rates are of the order of 0.2 MB/sec using general purpose computers. ChIPWig with random access only slightly degrades the performance and running time when compared to the standard mode. In the lossy mode, the average file sizes reduce by 2-fold compared to the lossless mode. Most importantly, near-optimal nonuniform quantization with respect to mean-square distortion does not affect peak calling and motif discovery results on the data tested.\n\nAvailability and ImplementationSource code and binaries freely available for download at https://github.com/vidarmehr/ChIPWig\n\nContactmilenkov@illinois.edu\n\nSupplementary informationIs available on bioRxiv.

bioinformatics↗