How to trim the adaptor sequence from the solexa small RNA sequencing data?

[satp]

Hi,

I have downloaded some solexa small RNA sequencing data from NCBI GEO(like this: http://www.ncbi.nlm.nih.gov/geo/quer...?acc=GSM273725). I want to trim the adaptor sequence from these sequences first, could anyone tell me which software can be used to do this work in personal computer effectively?

Leo

[regyre]

Hi,

If you're familiar with R and bioconductor, you can use some of the functionality from the ShortRead package : http://www.bioconductor.org/packages...ShortRead.html. Actually you should rather use R 2.9.0 (the development version) to get the latest version of the ShortRead package which has much more capabilities implemented.

There's been some emails with some code samples posted in this mailing list: https://stat.ethz.ch/mailman/listinf...sig-sequencing back in December-January.

Hope this helps,

N.

[satp]

Thanks for replying. But I don't famailiar with R and bioconductor and want to learn it in the future.

I wonder whether there is any software that is specifically developed for trimming the adaptor of deep-sequencing data.

[kmcarr]

If you are doing this as part of an alignment pipeline you could use Novoalign from Novocraft. (Check the sticky thread of software packages for a link.) I will trim the adapter on the fly. It's default screening sequence is the appropriate adapter for Illumina small RNA data but you can enter your own sequence as well.

I have also used the program fuzznuc which is part of the EMBOSS package. You provide a short sequence to screen for, which may include ambiguous codes and allowances for mismatches. The program does not trim the read, just outputs the locations where a match is found; you could then use this information with a perl script to actually trim the reads.

[satp]

Thanks for your advices. It seems that Novoalign can't be run in PC with small ram(my PC only has 2G ram). I haven't used Novoalign yet, could this software just trim the adaptor without aligning the sequence to the genome?

Just now I found that Shijun et.al.(http://www.ncbi.nlm.nih.gov/pubmed/18342361) suggested useing vectorstrip of EMBOSS package to trim the adaptor in his paper. I will try the fuzznuc and vectorstrip now.

Whether any other adaptor-trimming software exist?

[demis001]

I had the same problem six month ago. The easiest and fastest way is to write a script and remove and filter garbage reads from your data. To do that you need "adapter sequence information" from the guy who sequence your data.

You need the nc sequence of 3' and 5 prime adapter. Then you read each line and look for the adaptor using RegEx.

The most annoying part is, you will only find 1-2 million read out of 10 to 15 million containing either part of 3' or 5 pri adapter. A large protion of the reads does not contain adapter.

DD
Bioinformatics Research Analyst

[seq_GA]

Any one has tried trimming both 3' and 5' end adapter sequnces? Because we have a situation to find adapters in both 3' and 5' solexa reads.

I tried FASTX clipper, but it doesnot allow me to allow mismatches in the adapter sequence. Any better solutions?

$ fastx_clipper -h
	usage: fastx_clipper [-h] [-a ADAPTER] [-D] [-l N] [-n] [-d N] [-c] [-C] [-o] [-v] [-z] [-i INFILE] [-o OUTFILE]

Thanks.

[francois.sabot]

There is also cutadapt, allowing a % of distance..
in fastx_clipper, you can specify the -M value lower than your adapter size, for the bording mutation. Just remind that fastx_clipper remove one by one the adapter. Launch them in series using the pipe.

[satp]

Originally Posted by seq_GA

Any one has tried trimming both 3' and 5' end adapter sequnces? Because we have a situation to find adapters in both 3' and 5' solexa reads.

I tried FASTX clipper, but it doesnot allow me to allow mismatches in the adapter sequence. Any better solutions?

$ fastx_clipper -h
	usage: fastx_clipper [-h] [-a ADAPTER] [-D] [-l N] [-n] [-d N] [-c] [-C] [-o] [-v] [-z] [-i INFILE] [-o OUTFILE]

Thanks.

You can try vectorstrip in EMBOSS. It can trim both 3' and 5' end adapter sequences and allow mismatches in the adapter sequences.

[thinkRNA]

Originally Posted by satp

You can try vectorstrip in EMBOSS. It can trim both 3' and 5' end adapter sequences and allow mismatches in the adapter sequences.

Hi Satp,

I want to download vectorstrip so I can apply it to the Gigabytes of sequence data but I can only find websites with a graphical interface to the program, which is too slow and time-consuming for my large number of sequences.

do you know if I can download the program and execute it on the commandline?

Thank YOU so much.

[thinkRNA]

Never mind, I figured out that I need to install all of Emboss!

[mmartin]

Just in case anyone else stumbles over this thread, I'd like to repeat what francois said and point people to cutadapt (of which I'm the author). It does find adapters in the 5' and 3' ends of reads and allows mismatches.