PgmNr 277: Whole genome sequencing puts Cas9 off-target mutagenesis into the context of genetic drift.Authors:
L.M.J. Nutter 1; S. Khalouei 2; J.D. Heaney 3; D.G. Lanza 3; S.M. Murray 4; K. Peterson 4; J.R. Seavitt 3; J.A. Wood 5; A. Ramani 2
View Session Add to Schedule
1) The Centre for Phenogenomics, The Hospital for Sick Chlidren, Toronto, ON, Canada; 2) The Centre for Computational Medicine, The Hospital for Sick Children, Toronto, Canada, M5G 1X8; 3) Baylor College of Medicine, Houston, TX, 77030; 4) The Jackson Laboratory, Bar Harbor, ME, 04609; 5) Mouse Biology Program, University of California Davis, Davis, CA, 04609
There are reports demonstrating that Cas9 introduces off-target mutations and others that off-target mutation rates are low. The majority of reports investigate one or a few guide RNAs, which may result in sequence or chromosome location bias. The Knockout Mouse Phenotyping Project (KOMP2) produces mutant mice in a high-throughput pipeline using Cas9 for mutagenesis in the inbred C57BL/6N strain. This has enabled us to use whole genome sequencing to assess mutations in the genomes of 51 Cas9-derived founder mice representing 162 different gRNAs along with 25 inbred control mice. Illumina paired-end reads provided >35X coverage with ≥90% of bases with >25 reads. Variants (SNPs and indels) were identified using GATK4.0 and structural variants using an intersection of Lumpy, Manta, CNVkit, and Wham, followed by MetaSV. Variants were filtered out when they occurred in two or more mice, indicating the variant likely resulted from genetic drift rather than from Cas9 activity. We used CasOFFinder to identify predicted off-targets with up to 5 mismatches and one DNA or RNA bulge among the variants in the respective founder for each gRNA. There were 20 genes for which one or more Cas9-induced off-target mutations were identified (46 total with a range of 1-10 and average of 2.3 per founder). For 31 genes, no Cas9-induced off-target mutations were identified. Importantly, these analyses demonstrated that there was an average of ~3,500 variants unique to each animal – founder or untreated control. Two important conclusions can be drawn; (1) with appropriately designed Cas9 gRNAs off-target mutagenesis is rare; and (2) genetic drift within a carefully maintained inbred line of mice results in thousands of genetic variants between individuals within that line. These results have implications in the use of mice to model human disease, i.e. that backcrossing or outcrossing mice introduces significantly more variation than the use of Cas9 and the appropriate controls are littermate or line mate wild-type mice for most genetic experiments. These results also raise the question. What is “normal” genetic sequence in the context of model organisms and in humans - patients, controls, tissues and cell lines; for both assessing the specificity of Cas9 for genome editing and assessing the consequences of variants associated with disease?