"Concave" coverage pattern - problem with sample prep?

[mewahl]

Hi folks,

Our lab recently sent a bunch of yeast strains off for sequencing. I've noticed that the coverage pattern (reads per position vs. position on chromosome) is relatively flat for some of the strains, but is concave for others. The image below shows an example comparing the coverage pattern from one strain (with curvature, red) to another (without, green). The effect is more pronounced for long chromosomes than short ones.



I could imagine that if reads aligned more easily near the telomeres, this would produce a concave coverage pattern, but then all strains would show the concave pattern. Because the samples are affected differently, I don't think the problem is with the read alignment.

I could imagine that damage during gDNA collection would produce a convex curvature (due to loss of distal portions of chromosome arms) in damaged samples, but can't imagine any issue with sample preparation that explain a concave pattern.

Do any of you have any ideas what could cause this curvature in some samples but not others? Thanks in advance!

[ETHANol]

This is a wild guess but I would bet it has to be one of two things: 1) biases generated during PCR amplification of your library. Perhaps the GC content is different near the ends of the chromosomes. Or 2) The ends of the chromosomes fragment more efficiently.

[mewahl]

This is a wild guess but I would bet it has to be one of two things: 1) biases generated during PCR amplification of your library. Perhaps the GC content is different near the ends of the chromosomes. Or 2) The ends of the chromosomes fragment more efficiently.

Hmm, good idea...if there were subtle differences in the setup of individual PCRs, or in the fragmentation step for each sample, then the biases you mention could be showing up in some samples but not others. Thanks for your help!