Announcement

Collapse
No announcement yet.

Is there a replacement for wwwblast using new blastx+ toolkit

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Is there a replacement for wwwblast using new blastx+ toolkit

    Either my Google powers are weak, or there is no update to NCBI's wwwblast since they changed over to the blast+ toolkit, i.e. since 2004.

    So my question is, is there any replacement for the web interface functionality of wwwblast, that is, the graphical overview. It would be fairly easy to write a cgi wrapper for blast+, except the image of the blast hits would be missing.

    Or am I doomed to keep using the old wwwblast (which still works, of course).

  • #2
    I take it your using an old copy installed at your location.
    New stuff is available at : ftp://ftp.ncbi.nih.gov/blast/executables/LATEST/

    Comment


    • #3
      Our solution was to use the BLAST+ binaries via a local Galaxy server, see http://getgalaxy.org

      Currently only the core BLAST options are exposed in the wrapper (feedback welcome, ideally on the galaxy-devel mailing list), but by integrating BLAST+ into Galaxy you can do things like filtering the results and downstream analysis.

      This is particularly nice for BLAST'ing multiple query sequences at once, something wwwblast doesn't allow (IIRC).

      Comment


      • #4
        Originally posted by maubp View Post
        This is particularly nice for BLAST'ing multiple query sequences at once, something wwwblast doesn't allow (IIRC).
        Actually, you can blast multiple sequences at once with wwwblast, but thanks for the Galaxy suggestion.

        Comment


        • #5
          Originally posted by Richard Finney View Post
          I take it your using an old copy installed at your location.
          New stuff is available at : ftp://ftp.ncbi.nih.gov/blast/executables/LATEST/
          Thank you. But this link, the same that the NCBI web page has, does not lead directly to wwwblast, but if you instead go to ftp://ftp.ncbi.nih.gov/blast/executables/release/2.2.26 then there are recent versions of wwwblast. Still with the legacy toolkit, not blast+, but at least I now know that they are maintaining it.
          Last edited by dsenalik; 03-16-2012, 11:03 AM.

          Comment


          • #6
            Note that BLAST 2.2.26 is the final release of the legacy BLAST toolkit, from now on it will be BLAST+ only.

            Comment


            • #7
              Or try http://www.sequenceserver.com

              Comment


              • #8
                I found this thread looking for help with deploying wwwblast. I downloaded and installed the latest blast+ (2.2.27+), downloaded preformatted databases, and installed the latest wwwblast (2.2.26) on my 64bit RHEL box, and deployed Apache. Blast+ works from CLI, wwwblast is accessible both locally and remotely, but wwwblast does not seem to go further than just accepting data. I presume that wwwblast is just front end of actual blast+ and invokes installed programs as needed. But I am not clear how wwwblast is connected to blast+ programs as I could not find any explanation in instructions, which I simply followed. As wwwblast does not perform the intended job, it appears I have missed something important. Please advise.

                Comment


                • #9
                  wwwblast (i.e. blast.REAL) does not use any of the blast+ programs, it is completely standalone.
                  I have found the tricky part is configuring the file .ncbirc (which is a hidden file) and blast.rc because instead of error messages you may just get "no hits found"
                  There seems to be no example .ncbirc in the distribution, here is mine:

                  Code:
                  [NCBI]
                  Data=/usr/local/bin/blast/data
                  
                  [BLAST]
                  BLASTDB=/home/blastdbs

                  Comment


                  • #10
                    blast+ web interface

                    Thanks. NCBI helpdesk responded to my problem with suggestion that I have not enabled CGI execution in httpd.conf, so when I did it appeared to solve the problem. However, I am still getting a lot of "No hits found", so I guess something is still wrong. In this regard - where .ncbirc should be located? It is in my home directory (as user) - should it be where wwwblast execs are located? I have correct BLASTDB assignment in .ncbirc, but what is

                    Code:
                    [NCBI]
                    Data=/usr/local/bin/blast/data
                    for?

                    In more general sense, I asked NCBI help desk guy that responded if there is/will be ever a web interface for blast+ similar to what one gets when uses online access. The answer was "I am afraid not". Man, that sucks, huh? So I guess CLI is the only option to use local install of blast2.2.27+ suite - I have not looked at Sequenceserver yet, but I am going to.
                    Last edited by yaximik; 12-26-2012, 05:27 PM.

                    Comment


                    • #11
                      In my case it is in the same directory as the blast.REAL cgi program.
                      Code:
                      Data=/usr/local/bin/blast/data
                      in my case points to a directory with the five Matrix files:
                      PAM70 PAM30 BLOSUM80 BLOSUM62 BLOSUM45
                      I assume that works, but I admit I am not 100% sure about that.

                      Comment


                      • #12
                        Hi, based on your question "any replacement for the web interface functionality of wwwblast".
                        You may try:
                        a. SynaBlast - single query submission - http://synasite.mgrc.com.my/synasuit...enu.jsp?modN=1
                        b. SynaBlastMega - multiple query submission - http://synasite.mgrc.com.my/apps/syn..._BlastMega.jsp

                        Hope it helps

                        Comment


                        • #13
                          Ah, OK. In my case this data directory is in the same directory where all wwwblast executables and .cgi scripts are located, this is as installed following installation instructions. I should try to transfer .ncbirc to the same directory as this makes more sense. The blast installation directory has some ncbi_blast.rc file, which does not seem to have any meaningful role as its content makes little sense.

                          Comment


                          • #14
                            Originally posted by yannickwurm View Post
                            Well, I gave it a try as a front end to blast+. A data set of 16,000 sequences about 200 nt each, which could be entered only by means of copy/paste, was processed by blastx for about 48 hours. After blastx was done (the process ended) the output is flowing to stdout for 24 hours already without giving me a clue how long it is going to continue. The greyed out page in Firefox, from which the blastx job was launched, still shows data from the previous job. The worst thing is, as I found out, is that the output is not saved anywhere on machine or disk, so if I stop it or something stops it now, all results are lost. Even if I let it finish, I face an exciting opportunity to examine results of 16,000 queries manually in the browser window without any means to process data in a more efficient way. The interface, as it exists now, does not seem very usable. A little detail - when I posted this in google groups maintained by author(s), the message was simply deleted. Sorry.

                            Comment


                            • #15
                              Dear Yaximik,

                              apologies that SequenceServer isn't meeting your requirements. I think most users considering performing 16,000 BLAST queries will prefer using the command line over a graphical user interface.

                              Our aim with SequenceServer is to make BLASTing custom databases as accessible as possible to novice users. We cannot at this time provide all the features (e.g. in terms of output format & processing) that would satisfy power users and experienced bioinformaticians such as yourself.

                              Regarding your message to the mailing list - somehow it was autoclassified as spam by google groups but is now visible to all.

                              All the best & kind regards,

                              Yannick

                              Comment

                              Working...
                              X