Reader Comments

Post a new comment on this article

Interesting

Posted by CarolinB on 29 Jan 2011 at 09:56 GMT

Also, the average GC content for different read lengths was analysed. Results showed that GC content and obtained read length clearly correlate for both GS FLX and the Titanium platform. The applied sequencing methods did not only generate reads with different average levels of GC content, but also both pyrosequencing platforms show a significant decline in GC content once the read length exceeds a certain value (Figure 1).
http://plosone.org/article/info:doi/10.1371/journal.pone.0014519#article1.body1.sec3.sec1.p1

Can confirm this observation for an in-house metagenome dataset (454 GS FLX). However, I have yet to implement the correction process as described by the authors - so I can't tell how many sequences would be affected by this. Losing almost 30% of the data doesn't sound too good..

No competing interests declared.