Decipher Genomics · Ahmedabad, India

NGS pipelines, computational genomics, and the software underneath them.

Muhammed Murshid KP
NEXTFLOW DSL2·GATK4·DEEPVARIANT·CNVKIT·MANTA·PYTHON·QIIME2·DADA2·KRAKEN2·HUMANN3·SQLITE·GO·DOCKER·SINGULARITY·GROMACS·AUTODOCK VINA·STAR·DESEQ2·NEXTFLOW DSL2·GATK4·DEEPVARIANT·CNVKIT·MANTA·PYTHON·QIIME2·DADA2·KRAKEN2·HUMANN3·SQLITE·GO·DOCKER·SINGULARITY·GROMACS·AUTODOCK VINA·STAR·DESEQ2·
About

Bioinformatics engineer based in India, currently building clinical WES/WGS and genome-assembly pipelines at Decipher Genomics & Research. Work spans four domains: clinical genomics (germline and somatic variant calling, ACMG/AMP classification), genome assembly and metagenomics (de novo bacterial and eukaryotic assembly, 16S and shotgun profiling), predictive genomics (a multi-domain polygenic risk score platform), and structural bioinformatics (molecular dynamics, protein-ligand docking).

Also builds the systems software underneath the science: vcfilt, a zero-allocation VCF filter in Go that outperforms bcftools by 12.2x, published as a bioRxiv preprint, and the ClinVar classification-history tracker powering the tools section of this site.

Performance

12.2x faster than bcftools

vcfilt, a zero-allocation VCF filter written in Go, benchmarked against bcftools 1.18 on an 18GB VCF from 1000 Genomes chromosome 20.

vcfilt (plain VCF)
147,000 var/s
bcftools (plain VCF)
12,100 var/s
vcfilt (gzip VCF)
90,600 var/s
bcftools (gzip VCF)
11,500 var/s
Clinical Genomics

7,869 classification changes tracked

ClinVar Variant History, 2026. Ten monthly snapshots of the ACMG SF v3.3 panel (84 genes, 341,180 variants). Most changes reflect ClinVar’s aggregate label shifting as submissions arrive, not clinical reinterpretation.

84 genes341,000+ variants tracked7,869 classification changes10 months of monthly snapshots

Selected work

vcfilt — VCF streaming filter (Go)
12.2x bcftools, bioRxiv→
WES/WGS germline (trio-aware)
Nextflow DSL2→
ClinVar Variant Classification History
84-gene ACMG SF v3.3 panel→
Polygenic risk score platform
43 traits, SQLite→
De novo bacterial/yeast assembly
Nextflow DSL2→
16S rRNA amplicon profiling
QIIME2/DADA2→
Open to collaboration on genomics pipelines and tooling.Get in touch