Joining Contigs

[coldturkey]

Hi all,

after comparing a newbler assembly with a reference genome I have identified two contigs which contain the beginning and end (respectively) of the same gene. A PCR was run to bridge the gap and the product was sequenced yielding a 790bp fragment which overlaps both contigs
I am curious to know which tool is best to close the gap between the contigs?

regards

Brian

[jnfass]

Did you ask newbler to produce .ace files? -- I'd look at the ends of the contigs to see if there's any noticeable increase in read disagreement, or decrease in depth. It's possible that you simply had too few reads, and by chance did not span the gap ... or that other parts of your genome were similar enough to break the contig at that point. The .ace file should tell you the former, but you may need to run some blasts (or maybe pore through the alignment result files??) for the latter.

[sklages]

If you had told newbler to produce consed output, I'd use consed [1] for that
purpose. At least, that's what we are doing :-)
Most other software (except Staden's gap4 [2]) has very limited editing capabilities AFAIK.

cheers,
Sven

[1] = http://bozeman.mbt.washington.edu/consed/consed.html
[2] = http://sourceforge.net/projects/staden

[jkbonfield]

Gap4 works ok, but you need external tools (eg caftools from Sanger) to convert ace to gap4 as gap4's input format handling is shockingly bad still.

Gap5 is getting there. It's not really production ready, but a version exists (also at source forge) and I plan on making a new release again soon. The CVS tree (and hence next release) has full support for reading ACE files.

[coldturkey]

thanks everyone,

I did it in the end using Phrap. Although I had to make a quality file, on a temp basis, for the PCR product (quality of 20).

thanks again