Michal Hozza, Tomáš Vinař, Broňa Brejová.
How Big is That Genome? Estimating Genome Size and Coverage from k-mer Abundance Spectra.
In Costas S. Iliopoulos, Simon J. Puglisi, Emine Yilmaz, ed.,
String Processing and Information Retrieval (SPIRE),
9309 volume of Lecture Notes in Computer Science,
pp. 199-209, London, UK, September 2015. Springer.
Many practical algorithms for sequence alignment, genome assembly and other tasks represent a sequence as a set of k-mers. Here, we address the problems of estimating genome size and sequencing coverage from sequencing reads, without the need for sequence assembly. Our estimates are based on a histogram of k-mer abundance in the input set of sequencing reads and on probabilistic modeling of distribution of k-mer abundance based on parameters related to the coverage, error rate and repeat structure of the genome. Our method provides reliable estimates even at coverage as low as 0.5 or at error rates as high as 10%.