CRISPR Computing and Real-Time Continuous Molecular Recording Using Programmable Nucleic Acids

Abstract

Molecular computation has shown promise in the creation of programmable, responsive biological systems. This dissertation presents work in two areas involving molecular computation. In the first chapter, I describe my work on engineering conditional control of CRISPR-Cas9, a molecular tool popular for its easy programmability in targeting arbitrary DNA sequences by changing the corresponding sequence on its accompanying guide RNA. The engineered result, achieved by redesigning the guide RNA, is a Cas9 system that is conditionally active based on logical computations involving the particular nucleic acid sequences in its presence. I describe a unified design scheme for conditionally active and conditionally inactive Cas9 guide RNA and demonstrate its programmability in being able to sense and respond to sequence profiles in vitro based on Boolean logic (e.g. activate only if RNA sequences A and B are present), where the sensed sequences have no sequence dependence with the DNA sequence targeted by the guide RNA. In the second chapter, I present the implementation of a DNA-based molecular recorder capable of continuously recording concentration data over time in vitro. This molecular recorder uses DNA as both the agent sensing the concentration data as well as the recording substrate: concentration data are stored in sequential extensions of an ensemble of DNA strands recording in parallel over the same time interval. To read this information from the DNA ensemble, concentration data over time must be reconstructed by statistical inference. In this chapter I demonstrate the reconstruction of concentration data over time using a Markov chain Monte Carlo sampling algorithm, from DNA record sequences obtained via next-generation sequencing. I also present a more efficient inference strategy for reconstructing concentration data from long DNA records (resulting from long recording time and/or high concentration). Finally, I summarize these in vitro molecular computing results and discuss prospects for further expanding our current molecular computing capabilities.