Using false discovery rates to benchmark SNP-callers in next-generation sequencing projects
Nature Publishing Group
Open access. This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-nd/3.0/
Sequence alignments form the basis for many comparative and population genomic studies. Alignment tools provide a range of accuracies dependent on the divergence between the sequences and the alignment methods. Despite widespread use, there is no standard method for assessing the accuracy of a dataset and alignment strategy after resequencing. We present a framework and tool for determining the overall accuracies of an input read dataset, alignment and SNP-calling method providing an isolate in that dataset has a corresponding, or closely related reference sequence available. In addition to this tool for comparing False Discovery Rates (FDR), we include a method for determining homozygous and heterozygous positions from an alignment using binomial probabilities for an expected error rate. We benchmark this method against other SNP callers using our FDR method with three fungal genomes, finding that it was able achieve a high level of accuracy. These tools are available at http://cfdr.sourceforge.net/.
R.A.F. was funded by the Natural Environment Research Council (NERC). D.A.H. and M.C.F. were supported by the Wellcome Trust. No additional external funding received for this study.
This is the final version of the article. Available from Nature Publishing Group via the DOI in this record.
Vol. 3, article 1512
Place of publication