Skip to content
About
Software

Software

A selection of open-source software I have led, contributed to, or co-released alongside collaborators.

PolyLinkR

Prototype R package for gene-based pathway enrichment analysis, with explicit support for linkage disequilibrium between adjacent loci within pathways. Useful as evidence for polygenic selection when paired with selection-scan output (e.g. SweepFinder2). A modern and more efficient successor is in development by Raymond Tobler.

Code: github.com/ACAD-UofA/PolyLink Preprint: arxiv.org/abs/2004.03224

taxonomicfiltering

Workflow that uses Kraken2 metagenomic classification to filter contaminating sequences before mapping in ancient DNA pipelines. Up to ~94% reduction in mapping runtime with negligible loss of endogenous data; usable on personal machines. Released alongside Ravishankar et al., Briefings in Bioinformatics 26, bbae646 (2024).

Code: github.com/shyama-mama/taxonomicfiltering

cellular-constraint

Analysis code for the conserved facultative heterochromatin study identifying disease regulatory sequences. Released alongside Sinniah et al., Nucleic Acids Research 53(20) (2025).

Code: github.com/enakshi-sinniah/cellular-constraint

Adelaide Open Food Map

Interactive ML-driven map of Adelaide eateries that ranks “underrated” venues by combining Google Maps ratings with review counts. A side project, building on initial work by Lauren Leek.

Live tool: ysouilmi.com/adelaide_open_food_map Code: github.com/yassineS/adelaide_open_food_map

COSMOS

⚠︎
This project is no longer actively maintained. For new pipeline work I recommend Nextflow.

COSMOS is a Python library for building scientific pipelines that run on distributed computing clusters. It was originally designed for bioinformatics workflows but is general enough for any DAG-based distributed work, including image processing. The API uses plain Python rather than a domain-specific language, which I still consider a strength.

COSMOS was first published as an Application Note in Bioinformatics. Since then it has been re-written and open-sourced through a collaboration between The Lab for Personalized Medicine at Harvard Medical School, the Wall Lab at Stanford University, and Invitae, where it ran in production processing hundreds of thousands of clinical samples.

Code: github.com/Mizzou-CBMI/COSMOS2 Documentation: mizzou-cbmi.github.io/COSMOS2

Pipelines built on COSMOS

  • GenomeKey — GATK best-practices variant calling pipeline.
  • PV-Key — somatic tumour/normal variant calling pipeline.
  • MC-Key — multi-cloud implementation of GenomeKey.
Cloud Computing in ResearchFunding and Grants