Thomas Brooks headshot

Thomas G. Brooks


  • Research Associate
  • ITMAT Bioinformatics Lab
  • Institute for Translational Medicine and Therapeutics
  • University of Pennsylvania

I'm a research associate at the ITMAT Bioinformatics lab. I study transcriptomics and RNA-seq methodology, circadian biology and diurnal rhythms. Previously, I studied Riemannian geometry.

Research Interests

  • Transcriptomics, RNA-seq
  • Diurnal rhythms
  • Circadian biology
  • Methods development and benchmarking

Education

  • PostDoc in Bioinformatics, 2018-2023. University of Pennsylvania
  • PhD in mathematics, 2013-2018. University of Pennsylvania
  • BA in mathematics, 2009-2013. Cornell University

Projects

Temperature Rhythms

Skin temperature rhythms mark circadian rhythms in individuals, and are typically higher temperatures at night, opposite to the core body temperature rhythms. We use a measurement of wrist temperature that captures this variability in the UK Biobank cohort to show that individuals with poor rhythms end up with substantially worse health outcomes. View our results at the Temperature Biorhythm Atlas.

temperature rhythm profile

Nitecap - Circadian Omics Webapp

The transcriptome varies dramatically through the day. But these datasets are large and difficult to explore, so we made a web app that makes it easy to view, process, and share timeseries omics datasets. Try it out at nitecap.org.

Nitecap PCA plot example

RNA-Seq Simulation - BEERS 2

RNA-seq is a foundational technology in modern biology, but involves many subtleties that influence data processing. Reads do not correspond one-to-one to input RNA transcripts, but must be aligned, quantified, and normalized to generate results. Each of these steps involves difficulties. Through detailed simulation of the RNA-seq process, including library preparation, we can address these difficulties and benchmark the performance of analysis methods.

simulated RNA-seq coverage

dependentsimr

R package to simulate omics data with realistic correlation. Mimics the top PCA components of a dataset and is almost as simple and fast as generating entirely independent data.

See more projects on my github page.

Publications

* indicates equal contributions

Pre-prints