super low GC bias in many samples results in n/a iRep values

[willnotburn]

I have 120 MAGs across 25 samples. The MAGs are good quality: most are >90% complete with <5% contamination. But for most of these nice MAGs, there seems to be pretty much a logistic distribution for GC bias across the samples. The value is either roughly between 0.1 to 1 OR something like 2.2E-22. What is up with these super low GC bias values? What does it mean?

For any MAG, the super low GC bias for any sample results in failure to produce the corrected iRep values, even though all samples have the non-corrected "raw" indices of replication. Maybe I should just use those?

Much appreciate the help!

[christophertbrown]

Hi Paul, That sounds quite peculiar. Did you notice anything else that would suggest variation in sequencing quality across your samples? Given the extremes of GC bias, I would be skeptical of the uncorrected values. Chris

…

On Feb 23, 2020, at 4:27 PM, Paul Wilburn ***@***.***> wrote:  I have 120 MAGs across 25 samples. The MAGs are good quality: most are >90% complete with <5% contamination. But for most of these nice MAGs, there seems to be pretty much a logistic distribution for GC bias across the samples. The value is either roughly between 0.1 to 1 OR something like 2.2E-22. What is up with these super low GC bias values? What does it mean? For any MAG, the super low GC bias samples for any sample results in failure to produce the corrected iRep values, even though all samples have the non-corrected "raw" indices of replication. Maybe I should just use those? Much appreciate the help! — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe.