Saurabh Sinha

CellWHISPER infers direct cell-cell communication from spatial transcriptomics data, reducing errors by accounting for cell-type-specific expression and spatial organization.

MANTIS analyzes co-registered spatial metabolomics and transcriptomics data to discover metabolite patterns and gene-metabolite relationships with statistical rigor.

This study identifies ABCA1 as a cholesterol efflux protein that enhances anti-cancer functions of myeloid cells, with clinical relevance in breast tumors.

CellSP identifies and visualizes subcellular spatial patterns of mRNA, discovering functionally relevant modules in mouse brain and kidney cancer tissues.

The paper integrates metabolomics and single-cell transcriptomics to uncover gene-metabolite mechanisms driving osteoarthritis heterogeneity.

SpaceExpress detects phenotype-associated spatial gene expression differences using a neural network and spline regression across multiple spatial transcriptomics datasets.

The paper identifies non-coding DNA variations affecting P53 and cMYC-driven diseases by analyzing co-occupancy of these transcription factors and correlating DNA variations with gene expression changes.

The paper models sequence-to-expression relationships in ER-bound enhancers to identify functional variants and regulatory mechanisms in the estrogen receptor transcriptional program.

SPREd is a simulation-supervised neural network for gene regulatory network reconstruction that outperforms existing methods on synthetic and real yeast datasets.

SEAMoD models gene expression from enhancer sequences and identifies key transcription factors in neural stem cell differentiation.

The paper uses single-cell transcriptomics to show how genetic variants influence gene regulatory networks in honeybee brains, affecting aggression-related behaviors through specific cell populations and pathways.

BEDwARS improves bulk gene expression deconvolution by estimating cell type proportions and signatures simultaneously, showing robustness to noisy references and identifying mechanisms in a rare pediatric condition.

The paper compares thermodynamics-based models and a CNN called CoNSEPT for predicting enhancer function in Drosophila, finding that complex models can suggest interpretable mechanisms.

VarSAn identifies disease-relevant pathways from genomic variants using network analysis, showing complementary results to gene-based enrichment tests.

The paper identifies brain gene regulatory network plasticity as a mechanism underlying individual differences in honey bee behavior.

SERGIO simulates single-cell gene expression data using gene regulatory networks, producing datasets comparable to experimental data for benchmarking analysis tools.

The paper uses sequence-to-expression modeling to predict the impact of SNPs in an enhancer and validates mechanistic models with experimental data.

GENVISAGE identifies discriminative gene pairs in genomic data and visualizes their separation between biological classes with a 400X speedup over baselines.

This paper evaluates machine learning algorithms for predicting clinical drug response using preclinical samples and introduces TG-LASSO, which improves prediction accuracy for 7 out of 13 drugs.

This paper introduces GeneSet MAPR, a supervised learning algorithm for gene set analysis that leverages network patterns to improve identification of genes related to a given gene set, outperforming existing methods.

The paper develops a method to identify transcription factors regulating drug response by integrating genotype-phenotype data with TF binding predictions.

The paper introduces an information-theoretic framework to evaluate and design gene regulation experiments by quantifying information gain from perturbation experiments on D. melanogaster genes.

This study shows that regularization improves the generalizability of sequence-to-expression models by reducing overfitting in cross-validation experiments.

The paper identifies an EMAT gene signature that classifies breast cancer subtypes associated with distinct metastatic risks and improves prognosis accuracy.

This study introduces pGENMi, a probabilistic model that integrates multi-omics data to identify transcription factors regulating cell line response to cytotoxic treatment, finding 41% validation rate for predicted associations.

The paper identifies conserved neurogenomic response toolkits across species using multi-species transcriptomic analysis and systems-level functional grouping.

This paper evaluates cross-validation techniques for gene regulatory network models, finding that clustering-based CV provides more accurate generalizability estimates than random CV.

This study introduces InPheRNo, a method to reconstruct phenotype-relevant transcriptional regulatory networks, showing it can identify cancer-specific regulatory mechanisms and prognostic TFs.

The paper introduces ProGENI, a gene prioritization method that improves drug sensitivity prediction by integrating gene expression and prior biological knowledge, identifying new genes involved in chemoresistance.