Here are some of my past and ongoing projects, which I think are pretty cool. Not all projects are being actively developed, but I will certainly respond to issues and pull requests.

I am the main developer for (more recent projects are listed first):

• RootCauseDiscovery: Python and Julia package for detecting disease-causing genes in rare disease patients from gene expression data
• GhostKnockoffGWAS: Package for performing knockoff-based analysis for GWAS summary statistics data
• Ghostbasil.jl: (WIP) Provides Julia wrappers to the C++ code of ghostbasil
• Knockoffs.jl: Implements the knockoff filter framework for variable selection, which performs conditional independence testing and controls the FDR (false discovery rate)
• groupknockoffs: Simple app to solve group knockoff optimization, without Julia installed!
• EasyLD.jl: Julia utilities for downloading and reading LD (linkage disequilibrium) matrices stored in Hail's BlockMatrix format
• knockoffspy: A Python package that provides a direct wrapper over Knockoffs.jl
• knockoffsr: A R package that provides a direct wrapper over Knockoffs.jl
• MendelIHT.jl: Implements iterative hard thresholding (l0 penalized regression solver). It is highly optimized for handling compressed (binary PLINK) genotype data
• MendelImpute.jl: A package for genotype imputation, phasing, and (global/local) ancestry inference utilizing a reference haplotype panel. It is significantly faster than existing methods but slightly less accurate
• Thyrosim.jl: An updated version of THYROSIM, Thyrosim.jl produces individualized thyroid hormone predictions (TT4/TT3/TSH) based on a rather complicated ODE model
• VCFTools.jl: Julia utilities for handling VCF (Variant Call Format) files
• fastPHASE.jl: Julia wrapper for the famous fastPHASE genetics software. The primary use case is to allow the original program to run on binary PLINK data.

I am a contributor for (at least 5 commits):

• bge_analysis: Python, R, and Julia code for imputation and quality control scripts used for the blended genome-exome (BGE) data.
• QuasiCopula.jl: Implements a new class of distribution (Quasi-Copulas) that captures correlation among non-Gaussian random variables efficiently
• SnpArrays.jl: Julia package for handling binary PLINK formatted data. It has the fastest (matrix)-(vector) multiplication routine for compressed PLINK files as far as I know.
• MendelKinship.jl: Calculates various empirical and theoretical kinship coefficients, based on pedigree or genotype data.

I enjoy sharing my knowledge with others, so here are a few tutorials I made:

• Interfacing Julia with R/Python/C++ (as of early 2024)
• A general introduction to Julia (2022 version)
• A tutorial for multithreading and parallel computation in Julia.
• Some notes on random graph theory, presented in Biomath 203 at UCLA (2020)
• A tutorial to imputation and phasing using MendelImpute.jl presented at 2020 ASHG meeting (see homepage).
• A tutorial to iterative hard threhsolding (IHT) presented in the 2020 Lange Symposium.