Trimmomatic: Does it matter to trim the fastq before or after merging multiple replicates?

Question

[tunl]

For each sample, we have a group of replicates fastq files (which came from different lanes). To do analysis, we merge the multiple replicates fastq files into a single fastq file (via concatenation) for each sample.

I am now running Trimmomatic PE on the merged fastq pair for each sample. The steps performed are ILLUMINACLIP, LEADING, TRAILING, SLIDINGWINDOW, and MINLEN, following the example provided by the Trimmomatic online manual.

I am wondering if there is any difference between “trimming the original replicates fastq files before merging them” and “trimming the merged fastq file after the original replicates are merged”?

I may not fully understand the LEADING and TRAILING steps. The manual says:

LEADING: Remove low quality bases from the beginning. TRAILING: Remove low quality bases from the end.

These “from the beginning” and “from the end” make me worried: do they refer to the beginning and the end of a fastq file?

So in terms of performing LEADING and TRAILING steps, does it matter to run Trimmomatic on the fastq files before or after merging multiple replicates?

I’d greatly appreciate any advice.

Thank you very much!

Answer 1

[WouterDeCoster]

Leading and trailing reflect to the beginning and end of a read. As far as I know, trimmomatic processes each read individually and as such it shouldn't matter whether you first merge or first trim.

[Ming Tommy Tang]

I second this. I usually merge the fastqs together and then trim low quality bases and adaptors

[tunl]

Dear WouterDeCoster and tangming2005,

Many thanks to both of you!

Your information really helps!