Machine learning based prediction and analysis of anti-CRISPR proteins

Abstract

Anti-CRISPR (Acr) proteins are widespread amongst phage and promote phage infection by inactivating the bacterial host’s CRISPR-Cas defence system. Except for their universally short sequences, Acrs have little in common with each other. With very low sequence and structural similarity, at least 50 distinct Acr families have been identified across both bacterial and archaeal domains of life where they each use different molecular mechanisms to inhibit CRISPR-Cas systems. Outside the confined environment of a microbial cell, Acrs have inspired a number of downstream applications, from gene editing technologies and protein engineering to phage therapy, applications that are only limited by the relatively small number of known anti-CRISPR systems compared to the thousands hidden in sequenced genomes. In this talk, I will introduce our work in design and implementation of an all-in-one solution to better assist biologists to predict and analyze Acrs. This includes development of a novel machine learning based anti-CRISPR predictor (PaCRISPR) and a subsequent platform (AcrHub) to annotate known Acrs, predict novel Acrs and visualize the relationship between known and potential Acrs. These tools can either work independently or within the platform pipeline to facilitate prediction and downstream analysis of Acrs and thereby shorten the gap between prediction, functional characterisation, and eventual experimental validation.