Error and error mitigation in low-coverage genome assemblies.

Area of Science:

• Genomics
• Bioinformatics
• Evolutionary Biology

Background:

• Twenty-two new mammalian genome sequences are available, but most have low (2x) coverage.
• Low-coverage assemblies present challenges for accurate comparative genomics.

Purpose of the Study:

• To assess sequencing error rates in 2x mammalian genome assemblies.
• To evaluate the impact of sequencing errors on downstream analyses.
• To develop and test methods for sequencing error mitigation (SEM).

Main Methods:

• Comparison of 2x assemblies with high-quality ENCODE region sequences.
• Analysis of error sources, including base quality scores and read coverage.
• Development of automated SEM approaches utilizing error characteristics and cross-species comparisons.

Main Results:

• Sequencing error rate estimated at 1-4 errors per kilobase in 2x assemblies.
• Errors can be mistaken for true biological events (e.g., lineage-specific insertions).
• Automated SEM methods can mask or eliminate a substantial fraction of errors, reducing their impact on phylogenomic analyses.

Conclusions:

• Sequencing errors in low-coverage genomes can significantly distort comparative genomics results.
• Automated SEM provides a valuable tool for improving data quality, though additional sequencing is still recommended.
• Error-mitigated alignments are provided to facilitate downstream research.