Hello all,
I was curious what sort of estimate do people use when looking at RNA-seq experiments' alignment rates. I normally use tophat for such alignments.
Bowtie gives an unambiguous value for percent of aligned reads, while tophat does not offer any estimates. One can use percentage from bowtie log in "logs" folder - I believe that spliced reads should add maybe 10% on top of that. But still, it would be nice to have an exact metric.
Problem of course is that there are some reads that are aligned to more than one place, so just getting the number of reads in your final BAM file and dividing it by the number of reads in FASTQ file would not necessarily produce a meaningful results.
So, what do you do then? I have some ideas and in-house scripts, but I'm interested about what others do in this situations.
I was curious what sort of estimate do people use when looking at RNA-seq experiments' alignment rates. I normally use tophat for such alignments.
Bowtie gives an unambiguous value for percent of aligned reads, while tophat does not offer any estimates. One can use percentage from bowtie log in "logs" folder - I believe that spliced reads should add maybe 10% on top of that. But still, it would be nice to have an exact metric.
Problem of course is that there are some reads that are aligned to more than one place, so just getting the number of reads in your final BAM file and dividing it by the number of reads in FASTQ file would not necessarily produce a meaningful results.
So, what do you do then? I have some ideas and in-house scripts, but I'm interested about what others do in this situations.
Comment