extra high coverage regions in exome sequencing data

[lyz1030]

Hi everyone,

What reasons are the extra high coverage regions in exome sequencing data caused by? I want to know if can affect variant calling. What effect can it bring, false positive or false negative?

Thanks

[simonandrews]

There may be some variation in coverage due to changes in the capture efficiency of different regions - often attributed to GC content, but most capture system vendors are getting on top of this.

Larger variations in coverage are more likely to be the result of mismapping of repetitive regions. Depending on how you've done your mapping you may find that reads which can map to multiple positions are allowed to map more than once in the genome, which can produce towers of reads in your results.

Even if you only choose to see uniquely mapping reads you can still see overrepresented mismapped regions where you get reads from parts of the genome which are not present in the genome assembly, and are therefore invisible to the read mapping program.

In general you should be very suspicious of results coming from regions with very high read coverage compared to the rest of the sample since mismapped sequences will lead to false SNP predictions and would dilute any real SNPs in the region.