sequenceserver
sequenceserver copied to clipboard
wurmlab
Add DIAMOND/BWA support
DIAMOND is a fast blastx/blastp replacement http://www.nature.com/nmeth/journal/vaop/ncurrent/full/nmeth.3176.html
Maybe could output using --sam and then convert that to HTML with the reference and query sequences in some accessible (faidx?) format somehow? If this was done then using bwa instead of blastn would presumably be not much different.
Like #128, this is probably low priority.
yeah I had a look into this - unfortunately there is no plan for xml-type output; only a blast-table like if I remember correctly...
On 17 Feb 2015, at 06:23, Anurag Priyam [email protected] wrote:
DIAMOND is a fast blastx/blastp replacement http://www.nature.com/nmeth/journal/vaop/ncurrent/full/nmeth.3176.html
Maybe could output using --sam and then convert that to HTML with the reference and query sequences in some accessible (faidx?) format somehow? If this was done then using bwa instead of blastn would presumably be not much different.
Like #128, this is probably low priority.
— Reply to this email directly or view it on GitHub.
Presumably sam would be better than blast-table since the cigar string gives the actual alignment, but I agree the XML would have made things easier.
I had a chance to performance test diamond and it's ridiculously fast for large numbers of queries. For single queries, blastp is about equivalent. This would be a great feature for drag-n-dropping a file with lots of fasta entries though.
DIAMOND has had a lot of development recently, and is a tool that can produce BLAST compatible outputs. It is multiple times faster than BLAST - would you reconsider enabling Diamond for SequenceServer?
XML output generated by DIAMOND can now be imported and visualised in SequenceServer.
I am closing this issue. Please open a new one for BWA support if you still think is relevant.
Thanks.
I"m reopening this - as we want to be able run DIAMOND from within Seqserv (and it should now be possible to replace "like-for-like" (or almost))
Hello, I wanted to see if there were any updates on this feature? Looking to use a local seqserv instance with a large protein database, and this would be very useful. Thanks for your work on this tool!
Alternatively, would it be feasible to substitute the diamond blastp binary in the bin folder to use this as an alternative? I'm thinking this could be completed by:
- a small shell script that ran diamond blastp with default to outputting xml files
- Including a diamond formatted database with the same prefix as the blast database
This is a sort of temp fix, but I wanted to see if this was reasonable with the current underlying framework?
Hi @Rridley7 - very happy for you to explore this and potentially submit a pull request
Since version 2.0.9, DIAMOND can use BLAST databases: https://github.com/bbuchfink/diamond/releases/tag/v2.0.9 (with a small preparation step added in the next version: https://github.com/bbuchfink/diamond/releases/tag/v2.0.10) So it could be a faster, drop-in replacement for the BLAST binaries. Just the variables affecting memory usage will have to be tuned.
