RKRachit Kumar

Rachit Kumar

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Rachit Kumar has a total of 164 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    ARCHER-LD: Rapid Long-Range Linkage Disequilibrium Calculations at Biobank Scale using GPU Acceleration

    by Rachit Kumar, Pankhuri Singhal, David Zhang, Christopher Carson, Mitchell Conery, Alex Rodriguez, Tarak Nath Nandi, Marylyn D Ritchie, Mathialakan Thavappiragasam, Benjamin F Voight, et al.

    Summary: Linkage disequilibrium (LD) information from one’s own dataset is considered optimal for downstream analyses such as statistical fine-mapping. However, the computational complexity (≈N22 computations for N variants) lead most studies to use external reference panels, such as from 1000 Genomes. To capture LD across all variant pairs in biobank-scale whole-genome sequencing (WGS) datasets with hundreds of millions of variants, new computational strategies are essential. We present a novel approach that uses multi-GPU distributed computing to compute R2 for every variant pair in a dataset. On chromosome 22 (1.8 million variants) of the 30x WGS 1000 Genomes dataset, our method using eight consumer-level 12GB GPUs (NVIDIA RTX 2080TIs) takes 80 minutes, corresponding to a ≈8x speedup. We further demonstrate true biobank-scale performance in the Penn Medicine Biobank (PMBB; 57,170 samples), computing chromosome 1 LD (≈1.38 million variants) in 49 minutes versus 22.7 hours for PLINK, a ≈28x speedup. With this method, we successfully computed on the aforementioned 30x WGS 1000 Genomes dataset (≈120 million variants and ≈2500 samples) the entire LD matrix ($>$1e16, or 10 quadrillion elements) in under 6 hours using 512 NVIDIA 40GB A100 GPUs on the Department of Energy Argonne Leadership Computing Facility Polaris Supercomputer. We make this tool, coded in Python using CuPy, publicly available. Using this tool, researchers can leverage the full extent of their genomic data without relying on external LD reference panels and acquire more accurate, population-specific findings, particularly for groups underrepresented in existing databases. Availability and Implementation: ARCHER-LD is an open-source tool available on GitHub under the Apache License 2.0 at https://github.com/rachitk/ARCHER-LD. The version of record used in this manuscript is archived via Zenodo at https://doi.org/10.5281/zenodo.23399887. Supplementary Information: Supplementary material is available at Bioinformatics online.

  • Open Access

    ARCHER-LD: Rapid Long-Range Linkage Disequilibrium Calculations at Biobank Scale using GPU Acceleration

    by Rachit Kumar, Pankhuri Singhal, David Zhang, Christopher Carson, Mitchell Conery, Alex Rodriguez, Tarak Nath Nandi, Marylyn D. Ritchie, Mathialakan Thavappiragasam, Benjamin F. Voight, et al.

    Linkage disequilibrium (LD) information from one’s own dataset is considered optimal for downstream analyses such as statistical fine-mapping. However, the computational complexity ( computations for N variants) lead most studies to use external reference panels, such as from 1000 Genomes. To capture LD across all variant pairs in biobank-scale whole-genome sequencing (WGS) datasets with hundreds of millions of variants, new computational strategies are essential. We present a novel approach that uses multi-GPU distributed computing to compute R 2 for every variant pair in a dataset. On chromosome 22 (1.8 million variants) of the 30x WGS 1000 Genomes dataset, our method using eight consumer-level 12GB GPUs (NVIDIA RTX 2080TIs) takes 80 minutes, corresponding to a ≈8x speedup. We further demonstrate true biobank-scale performance in the Penn Medicine Biobank (PMBB; 57,170 samples), computing chromosome 1 LD (≈1.38 million variants) in 49 minutes versus 22.7 hours for PLINK, a ≈28x speedup. With this method, we successfully computed on the aforementioned 30x WGS 1000 Genomes dataset (≈120 million variants and ≈2500 samples) the entire LD matrix (>1e16, or 10 quadrillion elements) in under 6 hours using 512 NVIDIA 40GB A100 GPUs on the Department of Energy Argonne Leadership Computing Facility Polaris Supercomputer. We make this tool, coded in Python using CuPy, publicly available. Using this tool, researchers can leverage the full extent of their genomic data without relying on external LD reference panels and acquire more accurate, population-specific findings, particularly for groups underrepresented in existing databases. Key Messages: ARCHER-LD makes use of GPUs to speed up the computation of linkage disequilibrium dramatically. After conversion of genotype data to a disk-based matrix, we can compute LD in parallel over chunks of the genotype matrices for both phased and unphased data. By making use of a parallel formulation of LD calculations, ARCHER-LD can scale to any number of GPUs that are available. ARCHER-LD scales to real biobanks: in the Penn Medicine Biobank (57,170 samples) it computes whole-chromosome LD substantially faster than PLINK, with the speedup increasing as the dataset grows. ARCHER-LD will be publicly available for use in calculating in-sample LD and extension by other researchers.

  • Open Access

    PMBB Geno-Pheno Toolkit: A suite of scalable, reproducible pipelines for cross-biobank association analyses

    by Zachary B. Rodriguez, Lindsay Guare, Lannawill Caruth, Katie. M. Cardone, Christopher Carson, Tess Cherlin, Stephanie Mohammed, Hritvik Gupta, Rachit Kumar, Karl Keat, et al.

    Summary: Electronic health record (EHR)-linked biobanks generate unprecedented genomic and phenotypic datasets, but their scientific utility is constrained by data fragmentation across institutional silos and incompatible computing infrastructures, forcing researchers to rewrite ad-hoc scripts for each new environment. We present the PMBB Geno-Pheno Toolkit, a suite of modular Nextflow pipelines for biobank-scale association analyses. This note focuses on the toolkit’s SAIGE family of pipelines — supporting genome-wide (GWAS), exome-wide (ExWAS), and phenome-wide (PheWAS) association testing — together with the companion GWAMA and ExWAS meta-analysis pipelines that enable cross-biobank replication. All components are containerized (Docker/Apptainer) and orchestrated with Nextflow, allowing the same workflows to run unmodified on local HPC clusters, cloud platforms, and the All of Us Research Workbench. Complementary toolkit pipelines for PLINK-based GWAS, polygenic scoring, LD-based clumping, and phenotype harmonization are also available and briefly noted. Availability: The PMBB Geno-Pheno Toolkit is freely available at https://github.com/PMBB-Informatics-and-Genomics/pmbb-geno-pheno-toolkit under MIT open-source license.

  • Open Access

    AE-PocketMiner Uses Attention to Simultaneously Predict Cryptic Pockets and Their Allosteric Coupling

    by Si Zhang, Prajna Mishra, Devin Kelly, Rachit Kumar, Gregory R. Bowman

    Finding and targeting cryptic pockets could dramatically expand the druggable proteome. However, discovering these sites remains challenging since they are only open a fraction of the time. It is also difficult to predict the functional relevance of a cryptic site as this often requires insight into allostery. Here we introduce attention enabled (AE-)PocketMiner, an artificial intelligence (AI) method that uses a graph neural network with an attention mechanism to simultaneously predict the locations of cryptic pockets and their allosteric coupling to the rest of the protein from a single input structure. We show that AE-PocketMiner outperforms past methods for identifying cryptic pockets and recapitulates known allosteric interactions. Moreover, we experimentally confirm newly predicted cryptic pockets and mutations that allosterically control pocket opening. AE-PocketMiner thus provides a powerful framework for multiple steps of the drug discovery process—including pocket identification, prioritization, and assay design—that will help expand the druggable proteome.


  • Open Access

    SAIGE-GPU: accelerating genome- and phenome-wide association studies using GPUs

    by Alex Rodriguez, Youngdae Kim, Tarak Nath Nandi, Karl Keat, Rachit Kumar, Mitchell Conery, Rohan Bhukar, Molei Liu, John Hessington, Ketan Maheshwari, et al.

    Motivation: Genome-wide association studies (GWAS) at biobank scale are computationally intensive, especially for admixed populations requiring robust statistical models. SAIGE is a widely used method for generalized linear mixed-model GWAS but is limited by its CPU-based implementation, making phenome-wide association studies impractical for many research groups. Results: We developed SAIGE-GPU, a GPU-accelerated version of SAIGE that replaces CPU-intensive matrix operations with GPU-optimized kernels. The core innovation is distributing genetic relationship matrix calculations across GPUs and communication layers. Applied to 2068 phenotypes from 635 969 participants in the Million Veteran Program, including diverse and admixed populations, SAIGE-GPU achieved a 5-fold speedup in mixed model fitting on supercomputing infrastructure and cloud platforms. We further optimized the variant association testing step through multi-core and multi-trait parallelization. Deployed on Google Cloud Platform and Azure, the method provided substantial cost and time savings. Availability and implementation: Source code and binaries are available for download at https://github.com/saigegit/SAIGE/tree/SAIGE-GPU-1.3.3. A code snapshot is archived at Zenodo for reproducibility (DOI: [10.5281/zenodo.17642591]). SAIGE-GPU is available in a containerized format for use across HPC and cloud environments and is implemented in R/C++ and runs on Linux systems.


  • The Moment AI Arrived in the Clinic: Insights from the SAIL 2025 Year in Review

    by Pierre Elias, Kameron Collin Black, Payal Chandak, Elizabeth Healey, Rachit Kumar, Michelle M. Li, Hugo Morales, Brett Beaulieu Jones, Emily Alsentzer

  • Open Access

    CASTER-DTA: equivariant graph neural networks for predicting drug–target affinity

    by Rachit Kumar, Joseph D Romano, Marylyn D Ritchie

    Accurately determining the binding affinity of a ligand with a protein is important for drug design, development, and screening. With the advent of accessible protein structure prediction methods such as AlphaFold, predicted protein 3D structures are readily available; however, scalable methods for predicting binding affinity currently do not take full advantage of 3D protein information. Here, we present CASTER-DTA (Cross-Attention with Structural Target Equivariant Representations for Drug–Target Affinity), which uses an equivariant graph neural network (GNN) to learn more robust protein representations alongside a standard GNN to learn molecular representations to predict DTA. We augment these representations by incorporating an attention-based mechanism between protein residues and drug atoms to improve interpretability. We show that CASTER-DTA represents a state-of-the-art improvement on multiple benchmarks for predicting DTA, and that it generates novel insights for several related tasks. We then apply CASTER-DTA to create a large resource of the binding affinities of every drug approved by the U.S. Food and Drug Administration (FDA) against every protein in the human proteome and make these predictions freely available for download. We also make available a web server for researchers to apply a pretrained CASTER-DTA model for predicting binding affinities between arbitrary proteins and drugs.


  • Open Access

    Integrative multi-omics approaches identify molecular pathways and improve Alzheimer’s Disease risk prediction

    by Rasika Venkatesh, Katie M. Cardone, Yuki Bradford, Anni K. Moore, Rachit Kumar, Jason H. Moore, Li Shen, Dokyoon Kim, Marylyn D. Ritchie

    Alzheimer’s Disease (AD) is the most prevalent condition that impacts the aging population, with no effective treatment or singular underlying causal factor identified. As a complex disease, characterizing the genetic risk of developing AD has proven to be difficult; polygenic scores (PGS) exclusively use common variants which fail to fully capture disease heterogeneity. This study used univariate and multivariate approaches to characterize AD risk. Genome-, transcriptome-, and proteome-wide association studies (GWAS, TWAS, and PWAS) were conducted on 15,480 individuals from the Alzheimer’s Disease Sequencing Project (ADSP) R4 release to identify AD-associated signals, followed by pathway enrichment analysis. Integrative risk models (IRMs) were developed using genetically-regulated components of gene and protein expression and clinical covariates. Elastic-net logistic regression and random forest classifiers were evaluated using area under the receiver operating characteristic (AUROC), area under the precision-recall curve (AUPRC), F1-score, and balanced accuracy. These IRMs were compared against baseline PGS and covariate models. We identified 104 genomic, 319 transcriptomic, and 17 proteomic associations with AD under significant thresholds. Putatively novel associations were enriched in signaling, myeloid differentiation, and immune pathways. The best-performing IRM, random forest with transcriptomic and covariate features, achieved an AUROC of 0.703 and AUPRC of 0.622, significantly outperforming PGS and baseline models. Integrating univariate discovery approaches with multivariate modeling enhances AD risk prediction and offers insights into underlying biological processes.


  • Open Access

    CASTER-DTA: Equivariant Graph Neural Networks for Predicting Drug-Target Affinity

    by Rachit Kumar, Joseph D. Romano, Marylyn D. Ritchie

    Accurately determining the binding affinity of a ligand with a protein is important for drug design, development, and screening. With the advent of accessible protein structure prediction methods such as AlphaFold, predicted protein 3D structures are readily available; however, methods for predicting binding affinity currently do not take full advantage of 3D protein information. Here, we present CASTER-DTA (Cross-Attention with Structural Target Equivariant Representations for Drug-Target Affinity), which uses an equivariant graph neural network to learn more robust protein representations alongside a standard graph neural network to learn molecular representations to predict drug-target affinity. We augment these representations by incorporating an attention-based mechanism between protein residues and drug atoms to improve interpretability. We show that CASTER-DTA represents a state-of-the-art improvement on multiple benchmarks for predicting drug-target affinity and that it generates novel insights for several related tasks. We then apply CASTER-DTA to create a large resource of the binding affinities of every FDA-approved drug against every protein in the human proteome and make these predictions freely available for download. We also make available a web server for researchers to apply a pretrained CASTER-DTA model for predicting binding affinities between arbitrary proteins and drugs. Key Messages: CASTER-DTA is a novel architecture that makes use of equivariant graph neural networks to predict drug-target affinity, enabling rapid and interpretable predictions about how proteins and molecules bind to each other. Equivariant graph neural networks like those used in CASTER-DTA allow for more robust usage of protein 3D structural information and improve performance over non-equivariant approaches. CASTER-DTA includes a cross-attention layer to update protein residue and molecule atom embeddings that allows for interpretability about which residues are involved in the prediction. We can use this architecture for a variety of downstream tasks, and we have used CASTER-DTA to generate a resource of predicted binding affinities for every FDA-approved drug against every protein in the human proteome.


  • Open Access

    Genetic Algorithm Selection of Interacting Features (GASIF) for Selecting Biological Gene-Gene Interactions

    by Rachit Kumar, David Zhang, Marylyn DeRiggi Ritchie

    Feature interactions are particularly useful in modeling biological effects, such as gene-gene interactions, but are difficult to model due to the exponential increase in the feature space. We present GASIF, a Genetic Algorithm that selects features and their interactions for the purposes of solving a supervised classification problem, designed for the identification of gene-gene interactions. GASIF works by constructing individuals with a collection of chromosomes that represent a subset of features and their interactions. It then determines individual fitness as a combination of the number of unique features used and the cross-validation performance of a logistic regression classifier trained on that feature subset with an ElasticNet penalty. A variety of intuitive operations are used to select, mate, and mutate individuals from generation to generation to limit the search space of features and interactions. We evaluate this Genetic Algorithm on a real-world dataset of human brain transcriptomic data from neuropathologically normal postmortem samples and pathologically confirmed late-onset Alzheimer's disease individuals and determine the face validity of the gene-gene interactions that it identifies. Across multiple iterations of GASIF, we consistently identified the same features and interactions as most informative, all of which relate to genes known to be implicated in Alzheimer's disease.