[twu]

GSNAP can align bisulfite sequences, and was used to analyze methylation in the study of twins with multiple sclerosis that appeared in Nature last year. The latest version can handle both stranded and non-stranded bisulfite sequencing protocols.

Regards,

Tom

[rskr]

GSNAP can align bisulfite sequences, and was used to analyze methylation in the study of twins with multiple sclerosis that appeared in Nature last year. The latest version can handle both stranded and non-stranded bisulfite sequencing protocols.

Regards,

Tom

Last I checked the GSNAP paper's bench marking methodology, he only generated reads that could be mapped, with variants only in the middle portion of reads. I doubt it gets any better for bisulfite converted data.

[twu]

I did the benchmarking that way in the original GSNAP paper only to compare it against other programs that couldn't handle more complex reads, not to provide a full assessment of the capabilities of GSNAP. For example, I didn't benchmark its ability to find distant fusions, such as translocations, or its ability to find multiple indels and splicing in a single read. If you want to see an independent evaluation of GSNAP, there was a comparison done by researchers at University of Pennsylvania that appeared in Bioinformatics recently.

Regards,

Tom

[rskr]

I did the benchmarking that way in the original GSNAP paper only to compare it against other programs that couldn't handle more complex reads, not to provide a full assessment of the capabilities of GSNAP. For example, I didn't benchmark its ability to find distant fusions, such as translocations, or its ability to find multiple indels and splicing in a single read. If you want to see an independent evaluation of GSNAP, there was a comparison done by researchers at University of Pennsylvania that appeared in Bioinformatics recently.

Regards,

Tom

Did they evaluate GSNAP with bisulfite sequencing?

[twu]

No, it was a test of RNA-Seq simulated reads. And there are probably features of methylated reads that GSNAP does not assess, such as trying to evaluate various contexts, such as CG, CHG, and CHH. But it will align single-end or paired-end genomic reads having all possible substitutions of C to T with high speed and accuracy, including small and large indels and correctly accounting for T to C mismatches, and is not limited to variants in the middle of reads.

Regards,

Tom

[kalyankpy]

Hi Yuanxin,

I am quite interested in BSMAP. I am working on RRBS data from stem cells. I am trying to install BSMAP on my computer (Ubuntu 11.10, 64 bit). But it stops with an error. I posted the error messge at http://code.google.com/p/bsmap/issues/detail?id=7 . Can you fix this!

Kalyan

[yxi]

Hi Kalyan,

The error is in compiling samtools, bsmap used samtools API to read SAM files.

After google search I found someone experienced similar problems.

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=606004

Hope this could help

Thanks,

Yuanxin

Hi Yuanxin,

I am quite interested in BSMAP. I am working on RRBS data from stem cells. I am trying to install BSMAP on my computer (Ubuntu 11.10, 64 bit). But it stops with an error. I posted the error messge at http://code.google.com/p/bsmap/issues/detail?id=7 . Can you fix this!

Kalyan

[kalyankpy]

BSMAP installation on ubuntu

Hi Yuanxin,

I went through the link suggested by you! I noticed that there was a bug in the samtools in earlier versions which was rectified after 0.1.13 version (I noticed that the samtools you have provided with bsmap is 0.1.7).
I tried to overcome this bug by trying to download and install samtools-0.1.17 version from ubuntu repositories. I could install it successfully from repository. I also tried by downloading the sourcecode and compiled it with 'make' command. It workd perfectly. Assuming that problem is in the samtools, I replaced the samtools folder in bsmap with the newer version of samtools source code (0.1.18) and then tried compiling (with 'make').

This option took me few step ahead. Here I am attaching the output of the 'make' command. For me the problem doesnt seem to be in samtools rather it seems to be in some code that links BSMAP and SAMTOOLS (I am not a good programmer, I am learning! Its only my assumption. As a programmer you can analyze better!).

[yxi]

Good to know you solved it. We experienced a runtime error using the later versions of SAMTOOLS on Power7 CPU based system. So BSMAP included an early version SAMTOOLS.

Hi Yuanxin,

I went through the link suggested by you! I noticed that there was a bug in the samtools in earlier versions which was rectified after 0.1.13 version (I noticed that the samtools you have provided with bsmap is 0.1.7).
I tried to overcome this bug by trying to download and install samtools-0.1.17 version from ubuntu repositories. I could install it successfully from repository. I also tried by downloading the sourcecode and compiled it with 'make' command. It workd perfectly. Assuming that problem is in the samtools, I replaced the samtools folder in bsmap with the newer version of samtools source code (0.1.18) and then tried compiling (with 'make').

This option took me few step ahead. Here I am attaching the output of the 'make' command. For me the problem doesnt seem to be in samtools rather it seems to be in some code that links BSMAP and SAMTOOLS (I am not a good programmer, I am learning! Its only my assumption. As a programmer you can analyze better!).

[kalyankpy]

Hi, The problem still xists. I dont know if my message is clear for you. I could successfully install SAMTOOLS as a standalone application. However, I fail to install it when it is part of bsmap. Even if I replace the older version of SAMTOOLS with the newer version inside the BSMAP folder, the problem continues.

I liked the logic of your programme. I am just waiting for a solution to this installation problem.

Is it possible for you to send me the executibles for a 64 bit ubuntu system that can work for me (so that I dont need to compile them on mine as this problem persists)

[kalyankpy]

Fix to the ubuntu installation of BSMAP

Hi,

Yes finally the problem is solved. Here I explain the steps that help me resolve the installation steps in Ubuntu 11.04

tweak the makefile in the bsmap directory as follows

original line "$(CC) $(FLAGS) $(LIBS) $(REF_MODE) $(OLIGOLEN) $^ -o $@ $(THREAD) -lz -lbam"

Change it as "$(CC) $(FLAGS) $(LIBS) $(REF_MODE) $(OLIGOLEN) $^ -o $@ $(THREAD) -lbam -lz"

replace the older version of samtools source directory with the new version - 0.1.17

I dont understand why changing the order of "-lz -lbam" to "-lbam -lz" solves this issue.

[aniruddha.otago]

Hi Guys,

Recently we have published a paper in Nucleic Acids research- "Comparison of alignment software for genome-wide bisulphite sequence data". you might find it useful to have a look at it.. http://www.ncbi.nlm.nih.gov/pubmed/22344695

[rskr]

Hi Guys,

Recently we have published a paper in Nucleic Acids research- "Comparison of alignment software for genome-wide bisulphite sequence data". you might find it useful to have a look at it.. http://www.ncbi.nlm.nih.gov/pubmed/22344695

Doesn't look like you compared the accuracy of the mapping algorithms say by generating randomized cpg data. Any good reason why this wasn't done?

[yxibcm]

The BSMAP version (v1.02/v1.2) used in this paper is a bit old. You may want to use the latest version (v2.5) to get an up-to-date comparison.

Hi Guys,

Recently we have published a paper in Nucleic Acids research- "Comparison of alignment software for genome-wide bisulphite sequence data". you might find it useful to have a look at it.. http://www.ncbi.nlm.nih.gov/pubmed/22344695

[rskr]

The BSMAP version (v1.02/v1.2) used in this paper is a bit old. You may want to use the latest version (v2.5) to get an up-to-date comparison.

I guess old funky data "The accuracy went up when they trimmed it", has to get published somehow.