Artificial intelligence continues to squirm its way into many aspects of our lives. But what about biology, the study of life itself? AI can sift through hundreds of thousands of genome data points to identify potential new therapeutic targets. While these genomic insights may appear helpful, scientists aren鈥檛 sure how today鈥檚 AI models come to their conclusions in the first place. Now, a new system named SQUID arrives on the scene armed to pry open AI鈥檚 black box of murky internal logic.
SQUID, short for Surrogate Quantitative Interpretability for Deepnets, is a computational tool created by Laboratory(麻豆传媒社区) scientists. It鈥檚 designed to help interpret how AI models analyze the genome. Compared with other analysis tools, SQUID is more consistent, reduces background noise, and can lead to more accurate predictions about the effects of genetic mutations.
How does it work so much better? The key, LaboratoryAssistant Professor Peter Koo says, lies in SQUID鈥檚 specialized training. Koo explains:
鈥淭丑别 tools that people use to try to understand these models have been largely coming from other fields like computer vision or natural language processing. While they can be useful, they鈥檙e not optimal for genomics. What we did with SQUID was leverage decades of quantitative genetics knowledge to help us understand what these deep neural networks are learning.鈥�
SQUID works by first generating a library of over 100,000 variant DNA sequences. It then analyzes the library of mutations and their effects using a program called MAVE-NN (Multiplex Assays of Variant Effects Neural Network). This tool allows scientists to perform thousands of virtual experiments simultaneously. In effect, they can 鈥渇ish out鈥� the algorithms behind a given AI鈥檚 most accurate predictions. Their computational 鈥渃atch鈥� could set the stage for experiments that are more grounded in reality. LaboratoryAssociate Professor Justin Kinney, a co-principal investigator of the study, explains:
鈥淚n silico [virtual] experiments are no replacement for actual laboratory experiments. Nevertheless, they can be very informative. They can help scientists form hypotheses for how a particular region of the genome works or how a mutation might have a clinically relevant effect.鈥�
There are tons of AI models in the sea. More enter the waters each day. Koo, Kinney, and colleagues hope that SQUID will help scientists grab hold of those that best meet their specialized needs.
Though mapped, the human genome remains an incredibly challenging terrain. SQUID could help biologists navigate the field more effectively, bringing them closer to their findings鈥� true medical implications.
Written by: Luis Sandoval, Communications Specialist | sandova@cshl.edu | 516-367-6826
Funding
Simons Foundation, National Institutes of Health, Alfred P. Sloan Foundation
Citation
Seitz, E.E., et al., 鈥淚nterpreting cis-regulatory mechanisms from genomic deep neural networks using surrogate models鈥�, Nature Artificial Intelligence, June 21, 2024. DOI:
Core Facilites
鈥淭丑别 Sequencing Technologies and Analysis Shared Resource provides access to an array of high throughput Next Generation Sequencing (NGS) technologies. We offer cutting-edge technology alongside convenient in-house services for a broad range of genetic analysis.鈥� 鈥� Project Manager Sara Goodwin, Ph.D.