Biology subjects

Rajan, V.

Publications and source records attributed to Rajan, V..

15 recordsLinked to original sources

bioRxiv2025

Uncovering the Heterogeneity and Ontogeny of Mouse Thymic Macrophages Reveals an Unexpected Early Checkpoint Role

Thymic macrophages (TMs) maintain tissue homeostasis by clearing the large numbers of apoptotic cells generated during T cell development, but how TM heterogeneity relates to their developmental origin and role in thymocyte maturation remains incompletely understood. Using complementary flow-cytometric, single-cell transcriptomic, and genetic approaches, we resolved two major TM populations corresponding to TIMD4+ cortical and CX3CR1+ medullary/cortico-medullary macrophages. TIMD4+ VCAM1+ TMs displayed a prominent efferocytosis and apoptotic-cell-clearance program, whereas TIMD4- VCAM1+ TMs were enriched for antigen-presentation and interferon-response pathways. Fate mapping revealed unequal progenitor contributions to these populations, and CCR2 deficiency selectively reduced TIMD4- VCAM1+ TMs and thymic monocytes, supporting ongoing input from circulating precursors. Exploratory pseudotime analysis further identified a transcriptional continuum from Ly6c2+ Ccr2+ monocytes toward macrophage states. Using MaFIA fetal thymic organ cultures, AP20187-mediated depletion of Csf1r-expressing myeloid cells reduced CD4+CD8+ thymocyte differentiation and produced a coordinated accumulation of DN3 cells, loss of DN4 cells, and reduction in CD27 expression. These convergent changes identify the DN3-to-DN4 transition as a developmental stage that requires an intact Csf1r-expressing myeloid compartment, and establish a functional connection between the thymic myeloid niche and early {beta} T cell development. Together, our study refines the phenotypic and developmental organization of mouse TMs and reveals a previously under-appreciated requirement for myeloid-cell support during progression through the {beta}-selection checkpoint.

immunology↗

bioRxiv2025

PatternExtract: A facile, scalable pipeline for point pattern generation from spatial imaging data

We present PatternExtract, an open-source pipeline that generates accurate spatial point patterns from RGB pathology images and cell coordinate data without relying on composite channels or proprietary software. Using a novel two-kernel tissue segmentation method combined with automated pixel classification in QuPath, PatternExtract precisely excludes tissue artifacts such as necrosis and blood vessels to define spatial windows. Optimized on 568 diffuse large B-cell lymphoma images and validated on an independent cohort, the pipeline enables robust spatial analyses.

bioinformatics↗

bioRxiv2025

Interlinked roles for HEB and Id3 in fetal gamma-delta T cell commitment and functional programming

T cells expressing the {gamma}{delta} T cell receptor (TCR) develop in a stepwise process initiating at the {beta}/{gamma}{delta} T cell branchpoint followed by maturation and acquisition of effector functions, including the ability to produce interleukin-17 (IL-17) as {gamma}{delta}T17 cells. Previous studies linked TCR signal strength and fate choices to the transcriptional regulator HEB (Tcf12) and its antagonist, Id3, but how these factors regulate different stages of {gamma}{delta} T cell development has not been determined. We found that immature fetal {gamma}{delta}TCR+ cells from conditional Tcf12 knockout (HEB cKO) mice were defective in activating the {gamma}{delta}T17 program at an early stage, whereas Id3 deficient (Id3-KO) mice displayed a partial block in {gamma}{delta}T17 maturation and a defect in IL-17 production. We also found that HEB cKO mice failed to upregulate Id3 during {gamma}{delta}T17 development, whereas HEB overexpression elevated the levels of Id3 in collaboration with TCR signaling. Moreover, Egr2 and HEB were bound to several of the same regulatory sites on the Id3 gene locus in the context of early T cell development. Therefore, our findings reveal an interlinked sequence of events during which HEB and TCR signaling synergize to upregulate Id3, which enables maturation and acquisition of the {gamma}{delta}T17 effector program. One Sentence SummaryThe transcription factor HEB synergizes with TCR signaling to upregulate Id3, which is required for the maturation of fetal IL-17-producing {gamma}{delta} T cells.

immunology↗

bioRxiv2024

Pseudohypoxia-stabilized HIF2α transcriptionally inhibits MNRR1, a druggable target in MELAS

AbstractThe observation that amounts of the mitochondrial regulator MNRR1 (CHCHD2, AAG10, PARK22) are reduced in several pathologies, and that restoration of its level normalizes the pathological phenotype, prompted a search for compounds that could increase MNRR1 levels. High throughput screening of a 2400-compound drug and natural products library uncovered the antifungal drug nitazoxanide and its metabolite tizoxanide as effective enhancers of MNRR1 transcription. Using the mitochondrial disease MELAS (in which various mixtures, called heteroplasmy, of wild-type and mutant mitochondrial DNA (mtDNA) coexist) as a test, we showed that treating a cybrid MELAS model with tizoxanide could restore cellular respiration, enhance mitophagy, and, importantly, shift heteroplasmy toward more wild-type mtDNA. Furthermore, in MELAS patient fibroblasts, the compound could improve mitochondrial biogenesis, enhance autophagy, and protect the fibroblasts from LPS-induced inflammation. Chemical activation of MNRR1 is thus a potential strategy to improve mitochondrial deficits seen in MELAS. Investigation of the mechanism by which MNRR1 is reduced identified that two factors compete to regulate transcription at the MNRR1 promoter - RBPJ{kappa}, which stimulates it, and HIF2, which inhibits it. In MELAS cells there is a pseudohypoxic state that stabilizes HIF2, leading to transcriptional inhibition of MNRR1. Nitazoxanide reduces the levels of HIF2 by increasing the levels of PHD3, the prolyl hydroxylase that degrades HIF2.

molecular biology↗

bioRxiv2022

scMoMaT: Mosaic integration of single cell multi-omics matrices using matrix trifactorization

Single cell data integration methods aim to integrate cells across data batches and modalities, and obtain a comprehensive view of the cells. Single cell data integration tasks can be categorized into horizontal, vertical, diagonal, and mosaic integration, where mosaic integration is the most general and challenging case with few methods developed. We propose scMoMaT, a method that is able to integrate single cell multi-omics data under the mosaic integration scenario using matrix tri-factorization. During integration, scMoMaT is also able to uncover the cluster specific bio-markers across modalities. These multi-modal bio-markers are used to interpret and annotate the clusters to cell types. Moreover, scMoMaT can integrate cell batches with unequal cell type compositions. Applying scMoMaT to multiple real and simulated datasets demonstrated these features of scMoMaT and showed that scMoMaT has superior performance compared to existing methods. We also show that integrated cell embedding combined with learned bio-markers leads to cell type annotations of higher quality or resolution compared to their original annotations.

bioinformatics↗

bioRxiv2022

Neural Collective Matrix Factorization for Integrated Analysis of Heterogeneous Biomedical Data

MotivationIn many biomedical studies, there arises the need to integrate data from multiple directly or indirectly related sources. Collective matrix factorization (CMF) and its variants are models designed to collectively learn from arbitrary collections of matrices. The latent factors learnt are rich integrative representations that can be used in downstream tasks such as clustering or relation prediction with standard machine learning models. Previous CMF-based methods have numerous modeling limitations. They do not adequately capture complex non-linear interactions and do not explicitly model varying sparsity and noise levels in the inputs, and some cannot model inputs with multiple datatypes. These inadequacies limit their use on many biomedical datasets. ResultsTo address these limitations, we develop Neural Collective Matrix Factorization (NCMF), the first fully neural approach to CMF. We evaluate NCMF on two relation prediction tasks, gene-disease association prediction and adverse drug event prediction, using multiple datasets. In each case, data is obtained from heterogeneous publicly available databases, and used to learn representations to build predictive models. NCMF is found to outperform previous CMF-based methods and state-of-the-art graph embedding methods for representation learning in our experiments. Our experiments illustrate the versatility and efficacy of NCMF for seamless integration of heterogeneous data. Availabilityhttps://github.com/ncmfsrc/ncmf Contactvaibhav.rajan@nus.edu.sg

bioinformatics↗

bioRxiv2021

CRISPR Knock-in Designer: automatic oligonucleotide design software to introduce point mutations using CRISPR/Cas9

Knock-in of precise point mutations into protein-coding genes has been one of the earliest and most important applications of Clustered Regularly Interspaced Palindromic Repeats (CRISPR)/Cas9. The ability to perform such precise gene editing is crucial to interrogate the function of specific protein residues and to create models of human diseases caused by protein amino acid changes. The homologous protein residues can be mutated in model animal species, and the consequences of these mutations can be studied, leading to a better understanding of the disease in question. Design of point mutation knock-in strategies has been a combination of manual steps assisted by several computational tools resulting in a time-consuming process and preventing a single rapid and integrated solution. We have therefore designed CRISPR Knock-in Designer, which can perform rapid and automatic design of point mutation knock-in DNA oligonucleotides upon provision of the mutation, a guide RNA, and essential identifier or sequence information. The tool supports most experimentally established CRISPR types and has multiple options for the resulting oligonucleotides to satisfy the needs of most users. We also provide allele-specific PCR-based and restriction enzyme-based genotyping strategies as part of the program output. CRISPR Knock-in Designer adjusts to the genomic context of any target codon and tries to design knock-in strategies when a codon straddles two exons, a situation we explored in whole genomes of several model species. CRISPR Knock-in Designer output can also be adapted for use with some of the newer Prime Editing design tools to facilitate the introduction of a specific mutation sequence using this advanced technology.

genetics↗

bioRxiv2021

DruID: Personalized Drug Recommendations by Integrating Multiple Biomedical Databases for Cancer

Advances in next-generation sequencing technologies have led to the development of personalized genomic profiles in diagnostic panels that inform oncologists of alterations in clinically relevant genes. While targeted therapies for some alterations may be found, an effective therapeutic strategy should consider multiple and dependent genetic interactions that affect cancer progression, a task which remains challenging. There are ongoing efforts to profile cancer cells in-vitro, both to catalog their genomic information and study their sensitivity to various drugs. There is a need for tools that can interpret the personalized genomic profile of a patient in light of information from these biological and pre-clinical studies and recommend potentially useful drugs. To address this need, we develop a new algorithmic framework called DruID, to effectively combine drug efficacy predictions from a deep neural network model with information, such as drug sensitivity, drug-drug interactions and genetic dependencies, from multiple publicly available databases. We empirically evaluate DruID on cancer cell line data on which efficacy of many drugs have been experimentally determined. We find that DruID outperforms competing approaches and promises to be a useful tool in clinical decision-making.

bioinformatics↗

bioRxiv2020

ASTER: A Method to Predict Clinically Actionable Synthetic Lethal Interactions

A Synthetic Lethal (SL) interaction is a functional relationship between two genes or functional entities where the loss of either entity is viable but the loss of both is lethal. Such pairs can be used to develop targeted anticancer therapies with fewer side effects and reduced overtreatment. However, finding clinically actionable SL interactions remains challenging. Leveraging unified gene expression data of both disease-free and cancerous samples, we design a new technique based on statistical hypothesis testing, called ASTER, to identify SL pairs. We empirically find that the patterns of mutually exclusivity ASTER finds using genomic and transcriptomic data provides a strong signal of SL. For large-scale multiple hypothesis testing, we develop an extension called ASTER++ that can utilize additional input gene features within the hypothesis testing framework. Our extensive experiments demonstrate the efficacy of ASTER in identifying SL pairs with potential therapeutic benefits. CCS CONCEPTS* Applied computing [->] Computational genomics; Health informatics; * Mathematics of computing [->] Hypothesis testing and confidence interval computation. ACM Reference FormatHerty Liany, Anand Jeyasekharan, and Vaibhav Rajan. 2021. ASTER: A Method to Predict Clinically Actionable Synthetic Lethal Genetic Interactions. In Proceedings of ACM Conference. ACM, New York, NY, USA, 10 pages. https://doi.org/10.1145/nnnnnnn.nnnnnnn

bioinformatics↗

bioRxiv2020

Mining Pathway Associations from Networks of Mutual Exclusivity Interactions

Study of pairwise genetic interactions such as mutual exclusivity or synthetic lethality has led to the development of targeted anticancer therapies, and mining the network of such interactions is a common approach used to obtain deeper insights into the mechanism of cancer. A number of useful graph clustering-based tools exist to mine interaction networks. These tools find subgraphs or groups of genes wherein each gene belongs to a single subgraph. However, a gene may be present in multiple groups - for instance, a gene can be involved in multiple signalling pathways. We develop a new network mining algorithm, that does not impose this constraint and can provide a novel pathway-centric view. Our approach is based on finding edge-disjoint bipartite subgraphs of highest weights in an input network of genes, where edge weights indicate the significance of the interaction and each set of nodes in every bipartite subgraph is constrained to belong to a single pathway. This problem is NP-hard and we develop an Integer Linear Program to solve this problem. We evaluate our algorithm on breast and stomach cancer data. Our algorithm mines dense between-pathway interactions that are known to play important roles in cancer and are therapeutically actionable. Our algorithm complements existing network mining tools and can be useful to study the mutational landscape of cancer and inform therapy development.

bioinformatics↗

bioRxiv2019

Clustering using Automatic Differentiation: Misspecified models and High-Dimensional Data

We study two practically important cases of model based clustering using Gaussian Mixture Models: (1) when there is misspecification and (2) on high dimensional data, in the light of recent advances in Gradient Descent (GD) based optimization using Automatic Differentiation (AD). Our simulation studies show that EM has better clustering performance, measured by Adjusted Rand Index, compared to GD in cases of misspecification, whereas on high dimensional data GD outperforms EM. We observe that both with EM and GD there are many solutions with high likelihood but poor cluster interpretation. To address this problem we design a new penalty term for the likelihood based on the Kullback Leibler divergence between pairs of fitted components. Closed form expressions for the gradients of this penalized likelihood are difficult to derive but AD can be done effortlessly, illustrating the advantage of AD-based optimization. Extensions of this penalty for high dimensional data and for model selection are discussed. Numerical experiments on synthetic and real datasets demonstrate the efficacy of clustering using the proposed penalized likelihood approach.

bioinformatics↗

bioRxiv2019

Deep Augmented Multiview Clustering

We develop a deep learning based method to cluster arbitrary collections of matrices. Our method co-clusters each matrix in the input collection and also associates clusters across matrices thereby enabling discovery of cluster chains. We present preliminary findings on a clinical dataset.

bioinformatics↗

bioRxiv2019

Predicting Synthetic Lethal Interactions usingHeterogeneous Data Sources

MotivationA synthetic lethal (SL) interaction is a relationship between two functional entities where the loss of either one of the entities is viable but the loss of both entities is lethal to the cell. Such pairs can be used as drug targets in targeted anticancer therapies, and so, many methods have been developed to identify potential candidate SL pairs. However, these methods use only a subset of available data from multiple platforms, at genomic, epigenomic and transcriptomic levels; and hence are limited in their ability to learn from complex associations in heterogeneous data sources.\n\nResultsIn this paper we develop techniques that can seamlessly integrate multiple heterogeneous data sources to predict SL interactions. Our approach obtains latent representations by collective matrix factorization based techniques, which in turn are used for prediction through matrix completion. Our experiments, on a variety of biological datasets, illustrate the efficacy and versatility of our approach, that outperforms state-of-the-art methods for predicting SL interactions and can be used with heterogeneous data sources with minimal feature engineering.\n\nAvailabilitySoftware available at https://github.com/lianyh\n\nContactvaibhav.rajan@nus.edu.sg

bioinformatics↗

bioRxiv2019

Maximum Likelihood Reconstruction of Ancestral Networks by Integer Linear Programming

MotivationThe study of the evolutionary history of biological networks enables deep functional understanding of various bio-molecular processes. Network growth models, such as the Duplication-Mutation with Complementarity (DMC) model, provide a principled approach to characterizing the evolution of protein-protein interactions (PPI) based on duplication and divergence. Current methods for model-based ancestral network reconstruction primarily use greedy heuristics and yield sub-optimal solutions. ResultsWe present a new Integer Linear Programming (ILP) solution for maximum likelihood reconstruction of ancestral PPI networks using the DMC model. We prove the correctness of our solution that is designed to find the optimal solution. It can also use efficient heuristics from general-purpose ILP solvers to obtain multiple optimal and near-optimal solutions that may be useful in many applications. Experiments on synthetic data show that our ILP obtains solutions with higher likelihood than those from previous methods, and is robust to noise and model mismatch. We evaluate our algorithm on two real PPI networks, with proteins from the families of bZIP transcription factors and the Commander complex. On both the networks, solutions from our ILP have higher likelihood and are in better agreement with independent biological evidence from other studies. AvailabilityA Python implementation is available at https://bitbucket.org/cdal/. Contactvaibhav.rajan@nus.edu.sg

bioinformatics↗

bioRxiv2018

Copy number variants in clinical WGS: deployment and interpretation for rare and undiagnosed disease

PurposeCurrent diagnostic testing for genetic disorders involves serial use of specialized assays spanning multiple technologies. In principle, whole genome sequencing (WGS) has the potential to detect all genomic mutation types on a single platform and workflow. Here we sought to evaluate copy number variant (CNV) calling as part of a clinically accredited WGS test.\n\nMethodsUsing a depth-based copy number caller we performed analytical validation of CNV calling on a reference panel of 17 samples, compared the sensitivity of WGS-based variants to those from a clinical microarray, and set a bound on precision using orthogonal technologies. We developed a protocol for family-based analysis, annotation, filtering, visualization of WGS based CNV calls, and deployed this across a clinical cohort of 79 rare and undiagnosed cases.\n\nResultsWe found that CNV calls from WGS are at least as sensitive as those from microarrays, while only creating a modest increase in the number of variants interpreted (~10 CNVs per case). We identified clinically significant CNVs in 15% of the first 79 cases analyzed. This pipeline also enabled identification of cases of uniparental disomy (UPD) and a 50% mosaic trisomy 14. Directed analysis of some CNVs enabled break-point level resolution of genomic rearrangements and phasing of de-novo CNVs.\n\nConclusionRobust identification of CNVs by WGS is possible within a clinical testing environment, and further developments will bring improvements in resolution of smaller and more complex CNVs.

genomics↗