How to clean paired end reads?

[prs321]

I was on a roll while I was cleaning my data. I started with the first pair of the first sequence I was looking at. I applied cutadapt to remove the adapter sequence, then applied fastx trimmer. The results looked good on the FASTQC report, but I was still getting an error in the per sequence GC content.

The bigger issue that made me realize how badly I messed up was when I looked at the number of sequences present in my "cleaned" fastq file of the first pair. The number of sequences were less than before and didn't match up with the number of sequences in my second "uncleaned" pair.

So now I'm sitting here completely confused after thinking I was being productive and I was actually getting something done.

How exactly can I clean my data? I have 21 pairs to clean, each with sequences ranging from 35-255.

Also, how can I tell whether or not my paired end data overlaps with each other or not? Cause if they do not overlap then I was thinking about just merging the two files together for each specimen and then having only to analyze 21 files instead of 42 pairs.

[dpryan]

Here are some options for salvaging what you have without starting over. In the future, use something like trim_galore or trimmomatic to trim reads, since they can handle paired-end reads.

[prs321]

I just used trim_galore and it made my data a little bit better. No errors for the first pair end and the # of sequences match up.

1 issue. My second pair has an error for per base sequence quality.

If I were to trim off the bases that had bad quality scores, would that be fine? Such as if I used fastx_trimmer or is there a tool within trim_galore?

[dpryan]

Probably, it depends on exactly what the error looks like. See the -q option in trim_galore for quality trimming

[prs321]

For sure.

I went ahead and used the trimmer for my first pair. Both FASTQ files have no errors with a couple of warnings in each one.

Is it safe to just ignore these warnings considering the data looks pretty good? And also the fact that I used the tools to clean the data?