Photo of Michael Vinyard in front of a waterfall

Current: Building something new

Previous: PhD, Harvard University

Interests: deep learning, single-cell data, CRISPR mutagenesis, chromatin signaling

Software


Neural differential equations for single-cell dynamics

An intuitive bridge between AnnData and PyTorch

Neural differential equations

LARRY Dataset applied to ML evaluation workflows

Fetch and format single-cell matrices from AnnData

Easy, flexible, object-oriented RNNs

AnnData-facing wrapper of Spotify’s Annoy

PyTorch implementation of PBA

Other python libraries I’ve written:

  • ArchR-h5ad [ GitHub ]: convert .arrow files generated by ArchR to adata.h5ad

  • WebFiles [ GitHub ]: In-line python web downloader

  • AnnNMF [ GitHub ]: AnnData-compatible NMF for single-cell data

  • sgLib [ GitHub ]: sgRNA library design

  • seq-toolkit [ GitHub ]: sequence manipulation

  • genomeIO [ GitHub ]: genomics file parsing

  • Licorice [ GitHub ]: print pretty strings in python

  • GTFast [ GitHub ]: GTF access made quick and easy through strategic caching

  • MsigDB-adata [ GitHub ]: Interface scRNA-seq data with GSEA scoring

  • ABCParse [ GitHub ]: base class that comes with parsing batteries included

  • autodevice [ GitHub, Docs ]: in-line, device auto-detection for cross-platform compatibility

Projects


scDiffEq: neural differential equations for single-cell dynamics

SIMBA: Single-cell eMBedding Along with features

VR interface for exploring single-cell data

scATAC-seq benchmark

CRISPR-scanning of LSD1