gsMapper to SAM/BAM

[wjeck]

Is anyone aware of a conversion process to take gsMapper output (newbler) and turn it into the more standard SAM/BAM format?

[asifrim]

You can try using the bank2contig tool of the CVS version of the AMOS package. You can use the toAmos tool to convert the ace to an Amos bank which is the input of the bank2contig tool.

[bioinfosm]

I am confused. bank2contig gives sam format?

I am also interested in converting gsmapper output to sam format. any suggestions?

[asifrim]

The development version of bank2contig can take the -S option to output an AMOS bank as SAM. To get from gsMapper to an AMOS bank you'll have to use the outputted .ace file that gsMapper gives you. I haven't gotten it to to work properly though as the resulting SAM file doesn't seem to contain the base qualities.

I'm seriously considering scripting it myself if Roche doesn't deliver soon

[asifrim]

I've scripted a GS Mapper to SAM format converter in Ruby.
You can find it here:

http://homes.esat.kuleuven.be/~asifrim/gstosam/GstoSam.tgz

At the moment it only handles single end reads.
The code might be a bit ugly as I used this as an educational project to get into Ruby. Instructions can be found in the readme.

Caveat: The sam file compiles to bam, but I haven't thoroughly checked it for errors, so there might still be some bugs. Please feel free to report any if you see them

[MQ-BCBB]

Asifrim,

The bank2contig wtih the -S option doesn't give a SAM format. According to the documentation online, conversion to SAM should be using the -s option but as of version 2.0.8 that is not available. The -S option in version 2.0.8 give a Simple Layout Style.

I also tried your Ruby script GstoSam but it truncated prematurely.

MQ.

[asifrim]

Hi MQ,
I seem to have mistyped it, it should have have been downcase: -s for SAM
(http://sourceforge.net/apps/mediawik...le=Bank2contig)

Did the ruby script return any kind of error? What version of Ruby are you using? Are you outputting the standard GS mapper output or the tabbed version of the output? If I remember correctly the script works on the default output type (which is the multiline output per read).

[pmiguel]

I've scripted a GS Mapper to SAM format converter in Ruby.
You can find it here:

http://homes.esat.kuleuven.be/~asifrim/gstosam/GstoSam.tgz

At the moment it only handles single end reads.
The code might be a bit ugly as I used this as an educational project to get into Ruby. Instructions can be found in the readme.

Caveat: The sam file compiles to bam, but I haven't thoroughly checked it for errors, so there might still be some bugs. Please feel free to report any if you see them

My application: I would like to view the output of a 454 fragment sequence run mapped (resequence) back to the reference genome of the organism sequenced using IGV.

So I am using your script on gsMapper output. It seems to work to a first approximation. (Which is more than I can say for any of the other methods I have tried...) But it does not seem like I am quite there. Your script creates a .sam file -- but there is no header created. The sam specification claims http://samtools.sourceforge.net/SAM1.pdf

The SAM format consists of one header section and one alignment section.The whole header section can be absent, but keeping the header is recommended.

But attempts at converting this sam file to bam for sorting fail. For example:

 samtools view -b Synechocystis-1.sam
[bam_header_read] EOF marker is absent.
[main_samview] fail to read the header.

I am running a fairly recent version of samtools:

Version: 0.1.7 (r510)

Is there a reason why for my purposes a header section is required? Perhaps because I was running gsMapper, rather than gsAssembler?

--
Phillip

PS Yes, I agree Roche programs should offer SAM/BAM output as well as .ace