Invited: Identifying mutations from TILLING experiments

摘要

TILLING (Targeting Induced Local Lesions IN Genomes) is a reverse genetics approach to detect effects of globally induced mutations in a population and identify the individuals that have mutations in genes of interest. As long as the DNA sequence of the target gene is known and the organism of interest can be mutagenized, TILLING provides mutations in species where tools applicable to other model systems are unavailable. Importantly, organisms amenable to TILLING include both commercially valuable species such as rice, wheat, soybean, brassica, oat, and melon, and species important for research such as medaka, zebra fish, fruit flies, arabidopsis and nematodes. TILLING -by -Sequencing leverages next-generation sequencing and an overlapping pooled experimental design. It follows up the mutagenesis with deep sequencing of pools of individuals or populations of interest. Because of the high throughput of current sequencing technologies, deep sequencing to hundred and thousand fold coverage is possible [1]. Here I present on our method, Coverage Aware Mutation calling using Bayesian analysis, CAMBa, (read like the dance), which directly considers the pooled setup and sequencing coverage levels when calculating mutation and noise probabilities. Using data from two TILLING experiments, one with lower sequencing coverage variablility and data quality and the other with higher, we validate CAMBa's efficacy in identifying mutations, and demonstrate that it outperforms significantly other methods on sequence data of lower quality and higher variance in coverage across pools. We show that our method effectively discovers mutations in large populations with sensitivity of 92.5% and specificity of 99.8% [2].