Sai Zhang Profile picture
computational biology and machine learning. assistant professor at @UF. tsinghua (phd) → stanford (postdoc with @SnyderShot) → ufl (pi). views are my own.

Apr 16, 10 tweets

📢New preprint alert! Myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) is a complex disease that is affecting more people than we thought, yet it is significantly underrecognized and poorly studied. 1/n
medrxiv.org/content/10.110…

We systematically examined the genetics of ME/CFS using AI and WGS. We generated the largest WGS dataset to date for ME/CFS, based on 3 cohorts including @mecfsbiobank. We then developed HEAL2, a deep learning framework that predicts ME/CFS risk from rare coding variants. 2/n

@mecfsbiobank HEAL2 was derived from our previous model HEAL proposed a couple of years ago (), but significantly upgraded by incorporating PPIs using a GNN. We demonstrated HEAL2's superiority and constructed a very promising genetic risk score for ME/CFS. 3/n linkinghub.elsevier.com/retrieve/pii/S…

@mecfsbiobank Besides the predictive power, HEAL2 is also interpretable with an attention mechanism. 115 ME/CFS risk genes were identified from the model, showing significant LoF intolerance. 4/n

@mecfsbiobank We further dissected the functional heterogeneity of our candidate genes using bulk and single-cell RNA-seq and proteomic data, and revealed the functional significance of these genes across diverse tissues, systems, and cell types. 5/n

@mecfsbiobank Systems biology analysis also highlights the involvement of multiple systems in this diseases, particularly immune and neurological systems. 6/n

@mecfsbiobank To validate the disease relevance, we used multiple ME/CFS (sc)RNA-seq and proteomic datasets and demonstrated the down-regulation of our ME/CFS genes and module genes, spanning different immune cells. 7/n

@mecfsbiobank Lastly, we were interested in the genetic correlation between ME/CFS and other diseases. We examined this using both GWAS and rare variant analysis for >4,500 diseases. Interestingly, we found that ME/CFS is genetically correlated with depression and long COVID. 8/n

@mecfsbiobank Our study underscores the genetic, cellular and molecular complexity of ME/CFS and offers new insights into mechanisms. I hope this work will inspire more investigations into this "mysterious" disease. HEAL2 is a generic so it's readily to be used to study other diseases. 9/n

@mecfsbiobank More studies are in urgent need to push forward for effective diagnosis and treatment for ME/CFS! Immense thanks to all my collaborators in this study:@SnyderShot @DrMaureenHanson and others on the list and not on the list! Many thanks to all patients sharing their samples! 10/10

Share this Scrolly Tale with your friends.

A Scrolly Tale is a new way to read Twitter threads with a more visually immersive experience.
Discover more beautiful Scrolly Tales like this.

Keep scrolling