Skip to content
View s6aamani's full-sized avatar

Block or report s6aamani

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
s6aamani/README.md

Hi, I'm Aakash 👋

Computational Biologist & Bioinformatics Engineer
Statistical Genetics · Multi-Omics · Single-Cell Genomics · Scientific Software


🔬 About Me

I'm a postdoctoral computational biologist working at the intersection of statistical genetics, multi-omics integration, single-cell genomics, and bioinformatics software engineering.

My work ranges from developing R packages for network-based multi-omics integration to building large-scale genomic pipelines on HPC/cloud infrastructure and GPU-accelerated web applications for single-cell analysis.

I completed my PhD in Bioinformatics at Universität Bonn, where I worked on microbiome and multi-omics analyses of clinical dietary-intervention studies.

  • 🧬 Statistical genetics, GWAS/MAGMA, pQTLs and cross-ancestry genomics
  • 🧪 Multi-omics integration across transcriptomics, proteomics, metabolomics, microbiome and genetics
  • 🧠 Single-cell and single-nucleus RNA-seq analysis
  • ⚙️ Reproducible bioinformatics pipelines using Nextflow, Snakemake, Docker and SLURM
  • 💻 Scientific software development in R, Python and modern web frameworks
  • 📄 Recent work published in Nature Communications and Gut Microbes

🚀 Featured Projects

Full-stack, GPU-accelerated single-cell RNA-seq analysis platform for HPC environments.

Built with Next.js/React, FastAPI, RAPIDS/CUDA and SLURM, the platform runs an interactive single-cell workflow from QC through dimensionality reduction, clustering, differential expression and annotation on A100 GPUs.

The system includes dynamic GPU-worker orchestration, Seurat-to-AnnData conversion, large-dataset caching, interactive visualizations and an LLM-assisted analysis layer.


Contributor / Developer

Contributed to the continued development and modernization of hCoCena, an R framework for network-based horizontal integration and analysis of transcriptomic datasets.

The current package provides a Bioconductor-style workflow for correlation-network construction, module detection, multi-dataset integration, functional enrichment, cell-type analysis, longitudinal analysis and downstream biological interpretation.

My work has included development, debugging, performance and workflow improvements around the modernized package ecosystem.


Developer / Maintainer

Developed and maintain an R package implementing Vertical CoCena², extending network-based integration across heterogeneous omics layers.

The workflow constructs co-expression networks independently across layers and integrates them into shared multi-omic modules.

Engineering work includes:

  • R-package development
  • reproducible installation and dependency management
  • Docker-based execution
  • GitHub Actions CI
  • automated testing and regression checks
  • pkgdown documentation
  • reproducibility against published Vertical CoCena² analyses

SNPBOOST Proteogenomic Prediction

Built a high-throughput Nextflow/SLURM statistical-genetics pipeline to train genetic predictors for ~2,900 plasma proteins using UK Biobank Olink proteomics and genotype data.

The models are applied to ancient genomes and 1000 Genomes reference populations to investigate temporal and ancestry-related variation in genetically predicted protein abundance.

The project includes:

  • SNPBOOST model training
  • PLINK/PLINK2 scoring
  • ancient-genome and 1000 Genomes prediction
  • ancestry-adjusted regression
  • PCA and population-genetic analyses
  • scalable HPC execution
  • Flask/Plotly interactive result exploration

Cross-Ancestry Genomics — Million Veteran Program

Developed an R/SLURM analysis framework comparing European and African ancestry genetic architecture across Million Veteran Program traits.

The workflow integrates:

  • MAGMA gene-based association statistics
  • pathway enrichment
  • single-cell-informed cell-type associations
  • cross-ancestry concordance and divergence metrics
  • large-scale HPC processing

The analysis examines when genetic signals remain conserved across ancestries and when divergence emerges at gene, pathway and cell-type levels.


🧰 Tech Stack

Languages & Workflow

Python R Bash Nextflow Snakemake Docker SLURM

Genomics & Multi-Omics

GWAS Single Cell DESeq2 QIIME2 Multiomics Network Biology

Software & Infrastructure

FastAPI Next.js CUDA GitHub Actions DNAnexus


🔬 Research Areas

Statistical Genetics Multi-Omics Integration Single-Cell Genomics Transcriptomics Proteomics Microbiome Network Biology Scientific Software Machine Learning HPC


📄 Selected Publications

  • Klümpen, Mantri, et al. Cholesterol-lowering effects of oats induced by microbially produced phenolic metabolites in metabolic syndrome. Nature Communications (2026). DOI: 10.1038/s41467-026-68303-9

  • Klümpen, Mantri, et al. Calorie-restricted oat diet is associated with zonulin and short-chain fatty acid response in metabolic syndrome. Gut Microbes (2026). DOI: 10.1080/19490976.2026.2662687

  • TBK1/TNFRSF13B COVID-19 susceptibility study. npj Genomic Medicine (2021). DOI: 10.1038/s41525-021-00220-w


📫 Get in Touch

ORCID · omicsbuddy.com · aakashvmantri@gmail.com

Pinned Loading

  1. ancient-genomes-protein-prediction ancient-genomes-protein-prediction Public

    Genotype→protein prediction models trained on UK Biobank and applied to ancient European genomes, testing whether the proteome changed over the Holocene. GRM mixed models separate real selection fr…

    R 1

  2. cross-ancestry-magma cross-ancestry-magma Public

    Cross-ancestry concordance of MAGMA gene-based association signals across 39 Million Veteran Program traits, compared at gene, pathway and cell-type resolution.

    R 1

  3. hcocena hcocena Public

    Forked from BioCompNet/hcocena

    R 1

  4. BioCompNet/Vcocena BioCompNet/Vcocena Public

    Vertical omics integration

    R

  5. ebv-network-expansion ebv-network-expansion Public

    Python 1

  6. protein-prs protein-prs Public

    R 1