Partial alignment of Illumina reads.

PratikC · 02-23-2012, 01:17 AM

Hi all,
I have fastq files having Illumina paired end sequence reads. I want to align a part of read say read length is 75 bases and I want alignment in which at least 50 bases are aligned without any gap in between.

Why do I want this type of alignment is because there is a fusion of target sequence with reference sequence, so I want to ignore all completely aligned reads with reference and want to find out reads in which a part of target sequence is covered. (for example, as per above given calculation, out of 75 bases, I want to detect all reads having 25 bases from target sequence and 50 from reference sequence.

I tried with trimming the sequence and than aligning them separately but that is not working.

Is there any alignment tool which allows me to do alignment of 75 bases reads with parameters like align first 50 bases or last 50 bases with reference so that I can retrieve them and than align remaining bases with target sequence.?.?.?

Thanks in advance.

aurelielaugraud · 02-23-2012, 03:43 AM

you can try gem mapper, there is a utility called split mapper which can split the read to map both part of the spliced read independently .

PratikC · 02-23-2012, 08:47 AM

Hi aurelielaugraud,

Thanks for gem mapper.
In gem mapper I need to specify at what length sequence to be splited.

Actually what I need is not to specify the sequence length to be mapped. Because I dont know the exact fusion point in sequence reads. Out of total 75 base length, there might be fusion at 50 bp or 60 bp or whatever.

So what I need is an alignment with variable length, say one read is aligned with 50 bp match with reference and another read is aligned with 40 bp match and another with 60 bp match. (without any mismatch withing the matching stretch of sequence, thus whatever mismatch we get is in 5' end or 3' end)

In this way they all are something below 75 bp and thus I expect to find target sequence fused in the read at 5' or 3' end of the sequence.

aurelielaugraud · 02-23-2012, 12:16 PM

I am not sure how it works exactly but you can try every position ....???

HESmith · 02-23-2012, 01:37 PM

You can utilize split-end alignment, described here. For example, perform alignment of the first 25bp to your reference, then the last 25bp to your target, and get the union of those two sets. Repeat with first 25bp to target and last 25bp to reference.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Bowtie Illumina paired end reads alignment	empyrean	Bioinformatics	3	09-20-2011 09:51 AM
PubMed: Statistical Analyses of Next Generation Sequence Data: A Partial Overview.	Newsbot!	Literature Watch	0	11-30-2010 08:20 AM
Alignment Cover for Simulated Illumina Reads	BertieWooster	Bioinformatics	1	09-13-2010 07:15 PM
PubMed: Successful Pyrosequencing of GC-rich DNA Sequences by Partial Substitution of	Newsbot!	Literature Watch	0	05-27-2010 02:00 AM
entire and partial run	ptongyoo	General	0	04-22-2009 06:20 PM

02-23-2012, 01:17 AM	#1
PratikC Member Location: India Join Date: Sep 2011 Posts: 16	Partial alignment of Illumina reads. Hi all, I have fastq files having Illumina paired end sequence reads. I want to align a part of read say read length is 75 bases and I want alignment in which at least 50 bases are aligned without any gap in between. Why do I want this type of alignment is because there is a fusion of target sequence with reference sequence, so I want to ignore all completely aligned reads with reference and want to find out reads in which a part of target sequence is covered. (for example, as per above given calculation, out of 75 bases, I want to detect all reads having 25 bases from target sequence and 50 from reference sequence. I tried with trimming the sequence and than aligning them separately but that is not working. Is there any alignment tool which allows me to do alignment of 75 bases reads with parameters like align first 50 bases or last 50 bases with reference so that I can retrieve them and than align remaining bases with target sequence.?.?.? Thanks in advance.

02-23-2012, 03:43 AM	#2
aurelielaugraud Member Location: lyon Join Date: Aug 2009 Posts: 37	you can try gem mapper, there is a utility called split mapper which can split the read to map both part of the spliced read independently .

02-23-2012, 08:47 AM	#3
PratikC Member Location: India Join Date: Sep 2011 Posts: 16	Hi aurelielaugraud, Thanks for gem mapper. In gem mapper I need to specify at what length sequence to be splited. Actually what I need is not to specify the sequence length to be mapped. Because I dont know the exact fusion point in sequence reads. Out of total 75 base length, there might be fusion at 50 bp or 60 bp or whatever. So what I need is an alignment with variable length, say one read is aligned with 50 bp match with reference and another read is aligned with 40 bp match and another with 60 bp match. (without any mismatch withing the matching stretch of sequence, thus whatever mismatch we get is in 5' end or 3' end) In this way they all are something below 75 bp and thus I expect to find target sequence fused in the read at 5' or 3' end of the sequence.

02-23-2012, 12:16 PM	#4
aurelielaugraud Member Location: lyon Join Date: Aug 2009 Posts: 37	I am not sure how it works exactly but you can try every position ....???

02-23-2012, 01:37 PM	#5
HESmith Senior Member Location: Washington DC Join Date: Oct 2009 Posts: 280	You can utilize split-end alignment, described here. For example, perform alignment of the first 25bp to your reference, then the last 25bp to your target, and get the union of those two sets. Repeat with first 25bp to target and last 25bp to reference.