trimming Illumina data

[SDPA_Pet]

Hello, does anyone know the cutoff of trimming length. I have Illumina data from 2X150bp and 2X250bp sequencing. I'd like to trim them. I set the Q score to 30. How about length? What is the cutoff do you normally use?

For example, for 2X150bp data, should I discard all the reads short than 75bp or 100bp? How about 2X250 bp? I just want know the cutoff people usually use? Phred score >30 is normally use. I am not sure about the length.

Thanks

[Persistent LABS]

Hi SDPA_Pet,
The length cutoff will depend upon how much you gain while aligning the reads to genome [if your experiment is not a denovo assembly]. Smaller read lengths will increase the chance of alignments at multiple loci, which might not help you.
You can refer this publication: An Extensive Evaluation of Read Trimming Effects on Illumina NGS Data Analysis [http://journals.plos.org/plosone/art....pone.0085024]. The authors have used 70% of original read length as the length cutoff.

[GenoMax]

Originally Posted by SDPA_Pet

Hello, does anyone know the cutoff of trimming length. I have Illumina data from 2X150bp and 2X250bp sequencing. I'd like to trim them. I set the Q score to 30. How about length? What is the cutoff do you normally use?

For example, for 2X150bp data, should I discard all the reads short than 75bp or 100bp? How about 2X250 bp? I just want know the cutoff people usually use? Phred score >30 is normally use. I am not sure about the length.

Thanks

Q30 is an unnecessarily high cutoff, if all you are doing is aligning to a reference. You could omit trimming based on quality altogether in this case.

If you were going to do de novo assembly then you may want to trim at Q20-25.

If you are seeing a large amount of data getting trimmed then there may be some issue with your data that you would want to investigate further.

[SDPA_Pet]

Hi guys,

Yes, I am doing De novo metagenome assembly? Should I still use the length cutoff of 70%. For example, if I am trimming 150bp sequencing data, I will discard all the sequences shorter than 105bp.
Thanks.

[Persistent LABS]

Originally Posted by SDPA_Pet

Hi guys,

Yes, I am doing De novo metagenome assembly? Should I still use the length cutoff of 70%. For example, if I am trimming 150bp sequencing data, I will discard all the sequences shorter than 105bp.
Thanks.

Longer the reads, better will be the assembly. But of course you will be loosing some reads with high length cutoff. Important point is how much you are loosing. For example, if you are loosing only 5% reads, I think you are good to go ahead. Researchers have used even 20-30nt sequences to create draft genome assemblies [http://bib.oxfordjournals.org/content/11/5/457.full].