Project Description

Dr Dimitri Perrin

Queensland University of Technology

Will it cut? Predicting the efficiency of CRISPR-based gene editing

Data science and machine learning for bioinformatics

Thursday 4 July 2019

Dr Dimitri Perrin is Senior Lecturer at the Queensland University of Technology, where he leads the Biomedical Data Science group. His research interests are in developing new approaches to analyse, understand and optimise biomedical and social systems. His work therefore spans the areas of data science, modelling, computational biology and bioinformatics. Recent projects include gene editing (CRISPR), high-resolution biomedical imaging (CUBIC), and mobile apps for health research. Dimitri Perrin holds a Master’s Degree (Diplôme d’Ingénieur) from ISIMA and MSc from Université Blaise Pascal, and received his PhD from Dublin City University.

We are in the middle of a technological revolution: while methods to modify genomes have been around for some time, CRISPR provides a way to achieve this with unprecedented ease and precision. One of the most crucial parts in CRISPR experiments is the design of the “guide” sequence that will decide where the modification is made. This is not trivial, and there are a number of tools that aim to make this step more reliable. In this talk, we will discuss machine-learning approaches that try to identify efficient guides directly from their sequence. We will look at their individual performance, and at the surprising lack of agreement between tools. Consensus-based methods can partly address this, but limitations remain.