"Genetic genealogy-based approach enhances the discovery of trait associated loci in small-sample populations" from @bryandinh.bsky.social:

drive.google.com/file/d/1UY-P...

That's a wrap, #ASHG25!

"An evaluation of the prediction accuracy for polygenic score models across thirteen quantitative biomarkers and anthropometric measures in Native Hawaiians" from Kimberli Alatorre. Kimberli is a postbac in her first research experience!

drive.google.com/file/d/1t-rg...

and.. "Incorporating Dietary Information to Enhance Polygenic Prediction Models with Applications to Body Mass Index and Type 2 Diabetes" from Eunice Lee:

drive.google.com/file/d/1gU63...

First day of poster session is done! If you missed our poster but still want to check it out, you can use the link here!

"Pattern of genetic variation and enrichment of functional alleles in Native Hawaiians" from @jitang1024.bsky.social :

drive.google.com/file/d/1LWfy...

Thanks to the weather I've spent more time at O'Hare than #ASHG25 so far, but I'm now ready to go!

As usual, our research is well-represented at #ASHG2025, with four trainees presenting the advances in their research from this past year. Please go bombard them with your questions!

This manuscript is a part towards that goal, where we showed under-performance of PGS models for Native Hawaiians. Presence of admixture masked this somewhat, thus creating some disparity even within the population.

“I don’t think [my research is] something the New Republic would be interested in.”

How eerily ominous…
(No, I’m way behind on Star Wars shows… i don’t know what is going to happen. This scene just struck a nerve.)

#thisistheway

Bluesky now has over 10 million users, and I was #414,560!

As I turned on airplane mode going to the TAGC 2024 conference, I missed an email from the Dean... I was too cheap to pay $39.99 in-flight for internet so you can imagine the jitter for that 4-hr flight.

Always envious of those tenure-with-Death-Star-Lego photos.. did I go overboard here?

Happy Halloween! I’ve waited two years to do this - Mendel and his pea!

Apparently I’m also a cookie… in addition to being a city, mountain, candy, dance, and drosophila gene.

We applied the method to empirical data from Latinos and Native Hawaiians. We note results should be interpreted cautiously, as we did not have access to high quality sequencing data for these understudied admixed populations, and tree reconstruction methods can still improve!

We also benchmarked the methods on complicated models from stdpopsim, e.g. those with exponential growth and those with ancient samples from within-continent admixtures – gLike are all much more accurate and precise in general.

In simulations, gLike accurately estimated the demographic parameter of multi-way admixed populations using the true trees. The performance was a little worse when inferred trees (by tsdate) were used, but still appears to be much better than site-frequency-based methods.

We devised a method called gLike that derives the full likelihood of a genealogical tree. It uses a graph-based structure to trace each lineage in a tree through all possible trajectories of population memberships over time and computes the exact marginal probability.

Maybe a little late to the game, and don't know the norm practices on new platforms now... but if you are interested in the type of research we do, please check out our lab webpage!

chianglab.usc.edu

(We are looking for students, postdocs, and/or staff scientists too!)