Overall sequencing error rate estimation

[JanosMucsi]

Hi,

Given a Jellyfish output, I would like to estimate the overall sequencing error rate. We assume that the unique k-mers in this data set result from a single single-base error, and those are the only sequencing errors in the data set.
Any pointers? Thanks!

[Brian Bushnell]

You'll get the most accurate results by assembling and mapping to the assembly. But if you want to assume that an error spawns 31 error kmers, then count the total number of kmers with depth less than the bottom of the first trough, and divide by 31, and you'll get a rough approximation of the number of errors. To be more accurate, I would do this (using the BBMap package):

tadpole.sh in=reads.fq out=contigs.fa
bbmap.sh in=reads.fq mhist=mhist.txt ehist=ehist.txt