False Positives and False Negatives in Variant Calling

Errors in base calling may lead to the identification of non-existent variants (false positives) or to missing true variants in the sample (false negatives).

False Positive Rates

A high false positive rate increases the need to validate all variant discoveries with an orthogonal method, such as Sanger sequencing or Mendelian inheritance-based analysis.

False-Positive Rates from Runs on the SBL (version 3) and the Genome Analyzer Platforms as Identified by Sanger Sequencing

The above chart shows the estimated false positive rates for the sequencing by ligation (SBL) platform (version 3) and the Illumina Genome Analyzer, based on comparison to Sanger sequencing for four samples¹. Samples were composed of six genomic intervals spanning a total of 266 kb. The same samples were sequenced by Sanger sequencing, and the false positive rate was estimated as the percentage of variants called by each sequencing method that were not called as variants by Sanger sequencing.

The table above shows the estimated false positive rate for Illumina Genome Analyzer IIx² variant calls based on trio inheritance analysis. Of 64,381 assessable variant loci in the child (NA181506), 62,522 (97.11%) were attributable to Mendelian inheritance from the parents (NA18507 and NA18508), which corresponds to a false positive rate of 2.89%. While equivalent analysis from a SBL platform is not available (as of 10/27/2010), this Mendelian inheritance analysis of Illumina data demonstrates a very low false positive rate, as assessed by one of the more rigorous estimation methods.

Recreate these results or create comparable plots of your own data sets

1. Harismendy et al. (2009) Genome Biology 10:R32
2. Data from recent in-house runs on the Genome Analyzer IIx(2010), using TruSeq chemistry with 2 x 100 bp reads

Continue on to False Negative Rates Data