I wanted to bring this post from deep down in some thread up to start a new one, because I’m really interested in what people think about this. Should you disregard classification as a chimera for non-singleton sequences?
by laalaa99stl Â» Fri Feb 12, 2010 5:41 pm
I just had a thought (yes, yes, I know)…
If the exact same sequence is seen more than once, it really can’t be a chimera, can it? I mean, one really doesn’t expect the polymerase to fall off at exactly the same place over and over again, right? And because we typically use uniq.seqs before chimera.seqs, we may again be falsely boosting the positive count. But if the SNP density were low, I guess it would be impossible to pinpoint the precise breakpoint, so maybe throwing all the babies out with the bathwater is the right way to go. Or perhaps use a threshhold. If a I see a sequence 5 times, it’s not chimeric?