Towards integration of AI-enabled rationally designed de novo protein into solid-state chips

Abstract

For my dissertation in the UW Nanopore Laboratory, we developed methods aimed at creating permanent, robust, and integrated de novo protein sensors within an electronic SiNx chip. We identified critical design parameters needed for ideal de novo nanopore sensors with atomistic engineerability and reproducibility. This work integrates a wide array of de novo protein, with varying levels of stability, selectivity, and sequencing applicability, culminating in demonstrating ssDNA translocation using designed protein integrated into SiNx. Chapter 1 provides an introduction into nanopore sequencing, biological and solid-state nanopore systems, protein design through deep learning diffusion, and surface chemistry. Chapter 2 summarizes the work done in creating custom 3D printed microfluidics, control measurements for the hybrid system, CPD and ICR applications for surface coating analysis. Chapter 3 discusses the selection of protein from the early iterations of our design campaigns and identifies the primary parameters that are needed for stable insertions. Chapter 4 details improvements build off of the preliminary work in chapter 3 to demonstrate the translocation of small charged molecules of PAA and ssDNA. These results will provide insights towards rational design for hybrid sequencing platforms with stable integration of AI-enabled designer protein with solid-state platforms. With time, this work paves the way for atomically reproducible, purposefully engineered, and highly parallelized multi-omic hybrid nanopore system.