Extract specific reads from fastq or SRA

Question

[Ankit]

Hi everyone.,

I found one data of my interest from SRA. But the file is too big.

Is it possible to obtain specific reads from SRA (preferably) based on sequence string of my gene or region of interest?

Also if it is possible to do it with fastq file would be manageable?

Basically I want to avoid aligning full data which could be time-consuming and memory intensive. So I want to focus only reads of my interest.

Note:

1. I tried seqtk but it require read name and I don't know which read name to take.

2. grepping the sequence is another option but I would lose fastq based information

That's why I am looking for tools based on read subset/seed sequence. eg. AAATTCGC

I would appreciate any help.

Thank you

Answer 1

[GenoMax]

You can use bbduk.sh in filter mode providing sequence of your interest via literal= option. You will need the fastq file to be present.

[Ankit]

Thanks.

In case if somebody else want to do similar things, the command could be as follows:

/bbmap/bbduk.sh in=sample1.fastq out=unmatch_sample1.fastq outm=match_sample1.fastq literal=AGTTATTTTTATAGTGGAGAGAGATGGCGTTTAAGTGCAAATTTGTTAGTAGTTTTTT k=58

where k is the length of sequence

[GenoMax]

That k would get full length sequences. If you need to get partial matches then k should be set to less than 1/2 length of pattern one is searching for.