Pooja Agarwal L&S Biological Sciences
Understanding Deep Learning Models to Classify Variants in ASD
Although Autism Spectrum Disorder (ASD) has become better characterized in recent years, there is still much to learn about the genetics of ASD. Existing research has collected variants identified in ASD patients with novel mutations, but little is known about these variants’ effects. Previous research has identified variants in activation domains of transcription factors (TFs) that regulate gene expression. This project will focus on variants identified within activation domains of identified transcription factors. It is currently unknown how these variants impact activation domain activity and pathogenicity. My research aims to characterize the effect of these variants of unknown significance. To do this, I will investigate the function of existing deep-learning models that predict the activity of an activation domain based on sequence. I will then apply these models to the ASD variants to characterize their significance. This work will allow us to gain a deep knowledge of the impact of mutations in activation domains for ASD. In addition, it contributes to improving diagnosis for patients who carry novel mutations that have not been classified thus far.
Message To Sponsor
Thank you for supporting my project! Your donation is invaluable in helping me continue research I am passionate about this summer. Computational biology sits at the intersection of two fields I have long been interested in and is vital to the rapid advancement of biology. This fellowship and your funding have allowed me to continue to explore this area and work on understanding ASD patient variants, a project I hope will make a difference for patients. Thank you again for your generosity!