Speaker:
Title:
Sequence-to-function Machine Learning for Biological Sequences
Abstract:
Toehold switches, which are programmable nucleic acid sensors, face a design bottleneck; our limited understanding of how sequence impacts function often necessitates time-consuming screens to identify effective sensors. We employ machine learning approaches such as convolutional neural networks and language models to characterize and optimize toehold sequences in silico. We then broaden our scope to other biological sequences and develop an automated machine learning platform that mitigates technical challenges such as model design choices.