Paired end miRNA-data, trimming question

[Misa]

Hi!
I'm new on this forum and in this field of research. I've tried to find an answer to a probably very simple question, but have failed. So I try here and see where I end up?

I have paired end data for small RNA (sequenced on NextSeq500, library based on NEXTflex Small RNA-kit (Illumina compatible)). I realise most people have single end data, not paired end data for small RNAs. So I have problems with the adapter trimming. It's probably obvious how to handle that, but I'm new and I struggle...

I've used cutadapt, as recommended by Bioo Scientific, which works fine for R1, but not R2. I obviously don't understand which sequence I need to state for the adapter for the second read. Coz I had expected both R1 and R2 to be processed at more or less the same level since its so short reads? Any help?

cutadapt -a TGGAATTCTCGGGTGCCAAGG -A <adapter???> -o /proj/Trimming/B1_trimmed1_fwd.fq -p /proj/Trimming/B1_trimmed1_rev.fq --minimum-length 23 /proj/FastQ/B1_fwd.fastq.gz /proj/FastQ/B1_rev.fastq.gz

This is an example from when I ran the script, in this case I had stated the reverse complement for the adapter in the code above. I have also tried what results I get if I trim away the 5' adapter (but with the -g option instead) and don't get any wiser from that. *feel stupid*

Total read pairs processed: 8,827,513
Read 1 with adapter: 8,362,506 (94.7%)
Read 2 with adapter: 45,734 (0.5%)
Pairs that were too short: 91,927 (1.0%)
Pairs written (passing filters): 8,735,586 (99.0%)

Total basepairs processed: 1,341,781,976 bp
Read 1: 670,890,988 bp
Read 2: 670,890,988 bp
Total written (filtered): 950,264,383 bp (70.8%)
Read 1: 286,493,104 bp
Read 2: 663,771,279 bp

I appreciate any help and/or suggestions?
//Åsa

[Brian Bushnell]

You can use the BBMap package to find your adapter sequences like this:

bbmerge.sh in1=r1.fq.gz in2=r2.fq.gz outa=adapters.fa mininsert=17

Once you have the adapter sequences, you can trim like this:

bbduk.sh in1=r1.fq.gz in2=r2.fq.gz out1=trimmed1.fq.gz out2=trimmed2.fq.gz ref=adapters.fa tbo tpe ktrim=r k=21 mink=9 hdist=1

You should check the contents of adapters.fa to make sure it contains sequence. If you have a mix of adapters it may not work correctly and just output 'N' for the adapter sequences, so in that case you could simply use the included adapters in /bbmap/resources/adapters.fa. But using the specific adapter sequences for your reads is more precise.

[Misa]

You can use the BBMap package to find your adapter sequences like this:

bbmerge.sh in1=r1.fq.gz in2=r2.fq.gz outa=adapters.fa mininsert=17

Once you have the adapter sequences, you can trim like this:

bbduk.sh in1=r1.fq.gz in2=r2.fq.gz out1=trimmed1.fq.gz out2=trimmed2.fq.gz ref=adapters.fa tbo tpe ktrim=r k=21 mink=9 hdist=1

You should check the contents of adapters.fa to make sure it contains sequence. If you have a mix of adapters it may not work correctly and just output 'N' for the adapter sequences, so in that case you could simply use the included adapters in /bbmap/resources/adapters.fa. But using the specific adapter sequences for your reads is more precise.

I got a recommendation to use the following script, so I'm running that at the moment (first is the 3' adapter sequence, second is the reverse complement of the 5' adapter (I acctually had not tested that combination before )). Seems to do what I want so far:

cutadapt -a TGGAATTCTCGGGTGCCAAGG -A GATCGTCGGACTGTAGAACTCTGAAC -o R1.trim1.fq -p R2.trim1.fq R1.fastq.gz R2.fastq.gz

But thank you for your suggestion, I'll try that if this one does not work as I think it does.

[sergio]

If using NEXTFlex adapters you also nee to trim the first and last 4 bases from the adapter-clipped reads. Bioo adapters have four random bases at each end, so unless you remove them they will interfere with your mapping.

[yudibedi]

Late addition - but was running into the same issue with paired-end reads using NEXTFlex prep for small RNAs. These posts were really helpful.