Running Title: Error and Error Mitigation in Low-Coverage Genome Assemblies

The recent release of twenty-two new genome sequences has dramatically increased the data available for mammalian comparative genomics, but twenty of these new sequences are currently limited to ∼2x cov-erage. Here we examine the extent of sequencing error in these 2x assemblies, and its potential impact in downstream analyses. By comparing 2x assemblies with high-quality sequences from the ENCODE regions, we show that sequencing error, at 1–4 errors per kilobase, is sufficiently low for many purposes, yet still can have surprising effects. For example, an apparent lineage-specific insertion in a coding region is more likely to reflect sequencing error than a true biological event, and the length distribution of coding indels is strongly distorted by error. We find that the vast majority of errors are contributed by a small fraction of bases with low quality scores. Using a simple theoretical model, we show that error is strongly determined by regions of single-read coverage in the assembly, and in particular, by the ends of reads. We explore several approaches for automatic sequencing error mitigation (SEM), making use of the localized nature of sequencing error, the fact that it is well predicted by quality scores, and alignments with other species. While these methods cannot replace the need for additional sequencing, they allow substantial fractions of errors to be masked or eliminated at the cost of modest amounts of over-correction, and they can reduce the impact of error in downstream phylogenomic analyses. [Error-mitigated alignments of 32 vertebrate genomes can be downloaded a