Jonathan Golob, MD PhD

It’s midnight in America. And after just one year of Trump and GOP leadership, the country is weaker, more fearful, and more divided. Why would we ever want to keep going down this road?

Compared to one year ago, communities feel less safe and liberties are disappearing. Armed and masked federal agents are in American neighborhoods forcing their way into homes.

It’s midnight in America. Today, fewer men and women will go to work than a year ago, as jobs disappear despite record corporate profits.

@golob.bsky.social

In a highly collaborative effort with more than 50 authors, Dr. @golob & colleagues designed a @MarchofDimes DREAM Challenge aimed at leveraging longitudinal microbiome data & crowdsourcing, leading to 2 successful predictive models for preterm birth risk. marchofdimes.org/about/news/mar…

Thank you to all of our collaborators @golob @tomiko22 @AliceTang_mstp @AlennieRol @NimaAghaeepour @AdiTarca @gaiaandreoletti @iditkosti @PatriDiazGimeno @SirotaLab @DR_E_A_M and so many others as well as @MarchofDimes for support. Press release here: marchofdimes.org/about/news/mar…

We are so excited to share our work, a collaboration of over 50 authors led by @golob and @tomiko22 leveraging public microbiome data and crowdsourcing approaches to predict preterm birth, published in @CellRepMed today: cell.com/cell-reports-m… @UCSFPrecision @UCSF_BCHSI

@MarchofDimes @SirotaLab @msirota84 @tomiko22 @DR_E_A_M @umichmedicine @CellReports This was and is a thrilling project with wonderful and brilliant collaborators. I am genuinely optimistic and excited about the next steps, focused on getting this test to families and their care teams.

@MarchofDimes @SirotaLab @msirota84 @tomiko22 @DR_E_A_M @umichmedicine Our new approach for harmonizing together microbiome data from diverse studies is reported in a companion paper in Cell Reports Methods: pubmed.ncbi.nlm.nih.gov/37939711/ @CellReports

What makes a great AI/ML prediction in healthcare? I have four things I consider: 1. Is the prediction clinically meaningful and actionable? 2. Is the prediction robust? 3. Is the prediction generalizable? 4. Is the prediction reproducible?