miRNA fastqc sequence length distribution with UMI

Question

[maria2019]

I have single ended 75 bp miRNA reads (Quiagene miRNA kit) reads with UMI.

The fastqc report shows high peak at the 83-84 bp and illumina universal adaptor.

After removing the 5-3' adaptor ((5’-3’) AACTGTAGGCACCATCAAT) and also reads lower than 17bp with cutadapt, The sequence distribution peak is on 22-23.

I know that miRNA should be around 18-22 and UMI length 12. Doesn't it mean that I should see a peak around 30-34?

The code that I used was:

cutadapt -a AACTGTAGGCACCATCAAT --minimum-length 17 -o tri.fastq sample.fastq

Answer 1

[swbarnes2]

How did you process your fastq? bcl2fastq can be configured to remove the UMI from the read and put it in the read name; are you sure that wasn't done?

[maria2019]

I believe not. The head of fastq file is as follow:

@NB551007:45:HNKVLBGX5:1:11101:18335:1071 1:N:0:GCCAAT CTGGANGCGAGCCAACTGTAGGCACCATCAATNCCGTGCCCTCNAGATCGGAAGAGCACACGTCTGAACTCCAGTCACGCCAAT + AAAAA#EEEEEEEEEEEEEEEEEEEEEEEEEE#EAEEEEEEEE#EEEEEEEEEEEEEAEEEEEEEEEEEEEEEEEEEEEEEEEE @NB551007:45:HNKVLBGX5:1:11101:5844:1072 1:N:0:GCCAAT CTGTANGCACCATCAATCGACGTGAACAGAGATCGGAAGAGCACACGTCTGAACTCCAGTCACGCCAATATCTCGTATGCCGT + AAAAA#EEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEAEEEEEEEEEEEAEEE< AE/AEEEEEEEEEEEEE @NB551007:45:HNKVLBGX5:1:11101:23470:1072 1:N:0:GCCAAT CGTGGNGAGGAACAATTCTGAGAACTGTAGGCACCATCAATGAACTCGAACCCAGATCGGAAGAGCACACGTCTGAACTCCAGT + AAAAA#EEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEAEEEEEEEEEEEEEEEEEEE @NB551007:45:HNKVLBGX5:1:11101:12496:1074 1:N:0:GCCAAT TCGCTNCGATCTATTGAAAGTCGGCCCTCGACACAAGGGTTTGTAACTGTAGGCACCATCAATTCCCTTATTGCCAGATCGGAA + AAAAA#EEAEAEEEEE6EEEEE/EEEEEEEEEEEEEEEEEEEAEEEEEEEEEEEEEEEEEEEEEEAEEEEEEEEEEEEEEEEEE

In a downstream analysis I want to use UMI-tools for deduplication. I should actually have the UMI name on the read name to be able to work on it. I searched and looks like I can use fastp to remove the UMI from the read and move it to the read name.

Now my question would be once I have done that, for the trimming with cutadapt, should I remove reads higher than say 40 bp? Just keep 17-40 reads?