The Data Driver: How Tina Hernandez-Boussard is Shaping Inclusive Health Care (SOM Annual Report): “Data was my ticket to a different world,” she recalls.
Read the story here:
The Shah Group has a new systematic review — https://jamanetwork.com/journals/jama/fullarticle/2825147 — in JAMA on Testing and Evaluation of Health Care Applications of Large Language Models, which the editor has chosen to make free.
The review categorizes 500+ papers based on the data used in the evaluation, the health care task performed, the linguistic task evaluated, the dimensions of evaluation, and the medical specialty that did the study.
Congratulations, Shah Group!
| Stanford Medicine researchers, after creating an AI-based algorithm to find complex structural variants in the human genome, learned those variants likely contribute to psychiatric disease. |
The 3 billion base pairs that constitute the human genome — the matching jigsaw puzzle pieces of adenine pairing with thymine and cytosine pairing with guanine — are not just the body’s instruction manual. Rearrangements in the order of those base pairs are markers of the origins of disease and of our evolutionary history. They can be simple, when a handful of base pairs switch places. They can also be complex, such as when a stretch of tens of thousands of base pairs inverts and is missing multiple sections.
Read the story here: https://med.stanford.edu/ne
RNA secondary and tertiary structure is critically involved in ribozyme and ribosomal rRNA function, as well as viral and cellular regulation. Traditional experimental methods for RNA structure determination such as X-ray crystallography or chemical mapping are incisive; however, these approaches suffer from low-throughput and low-dimensionality, respectively. Computational approaches, leveraging evolutionary signals from correlated positions’ mutations, provide an alternative means to infer RNA structures. However, these methods require assembly, and face challenges due to statistical biases inherent in multiple sequence alignment (MSA). Furthermore, these methods cannot make use of the full spectrum of natural variations seen for a given RNA element.
Read it here:
https://www.biorxiv.org/content/10.1101/2024.10.03.616574v1