NGS: alignment, variant calling, and filtering Erik Garrison Bioinformatics for @daptation Genomics Weggis, Switzerland March 3, 2015 Overview 1. Genesis of variation (SNPs and indels) 2. Causes of sequencing error 3. Sequence alignment 4. Alignment-based variant detection 5. Assembly-based variant detection 6. Haplotype-based variant detection 7. Primary filtering: Bayesian callers 8. Post-call filtering: Hard filters, SVM DNA DNA “twisted” view of ladder A SNP A point mutation in which one base is swapped for another. AATTAGCCATTA AATTAGTCATTA (some) causes of SNPs ● Deamination ○ cytosine → uracil ○ 5-methylcytosine → thymine ○ guanine → xanthine (mispairs to A-T bp) ○ adenine → hypoxanthine (mispairs to G-C bp) Deamination of Cytosine to Uracil http://en.wikipedia.org/wiki/Deamination (some) causes of SNPs ● Depurination ○ purines are cleaved from DNA sugar backbone (5000/cell/day, pyrimidines at much lower rate) ○ Base excision repair (BEP) can fail → mutation Multi-base events (MNPs) ● MNPs ○ thymine dimerization (UV induced) ○ other (e.g. oxidative stress induced) http://www.rcsb.org/pdb/101/motm.do?momID=91 Transitions and transversions In general transitions are 2-3 times more common than transversions. (But this depends on biological context.) https://upload.wikimedia.org/wikipedia/commons/3/35/Transitions-transversions-v3.png An INDEL A mutation that results from the gain or loss of sequence. AATTAGCCATTA AATTA--CATTA