Matthew Berginski's Personal Webpage

I am a computational biologist at Tempus Labs, where I specialize in using machine learning and other computational techniques to analyze, understand and use biological data.

Previously, I was a Research Scientist in the Pharmacology department at the University of North Carolina and I worked in Shawn Gomez's lab and Gary Johnson's lab. I specialized in bioinformatics with a focus on image analysis, machine learning and quantification of high-throughput data. I also worked with the Bioinformatics and Analytics Research Collaborative (BARC) a bioinformatics core facility at UNC.

I do most of my work using R, but have also written code using MATLAB, Perl and Python. You can find a listing of my public projects, publications and open source code here:

Public Projects

• The Focal Adhesion Analysis Server: A web-based image processing tool dedicated to processing images of focal adhesions. Focal adhesions are a sub-cellular structure that allow individual cells to connect to the environment. This tool has been used to process over 2,000,000 images. Source code available through github.
• The Dark Kinome Knowledgebase: A website for the disemination of results related to understudied kinases from an NIH-funded consortium. Source code available through github.
  • The Dark Kinase Expression Browser: A data browser dedicated understanding understudied kinase expression through data from the GTex and HPM consortiums. The application was developed using the R shiny framework. Source code is available through github.

• Shiny Applications: websites based around the R shiny framework that I've been part of building.
• CORAL: Visualization application developed in Phanstiel lab for showing kinase-centric data in several forms (primary helped with deployment and bug testing). Github
• Embryonic Expression Transcriptomics: Transcriptomics data browser developed in collaboration with Parnell Lab. Github
• EVD68 Spatial Transcripts: Visualization tool for a spatial transcriptomics data set developed with the Vogt Lab. Github

Preview Applications

• Kinotype Viability Predictor: A shiny framework based tool that accepts RNAseq results and then predicts the effect on cell viability of ~230 kinase inhibitors. github

Source code for the other projects I've worked on are available through github.

A list of my publications is available through google scholar.

More details about my work history is available through LinkedIn.