Loading ...
Sorry, an error occurred while loading the content.

Error running latest ssaha

Expand Messages
  • sraxylan
    I m trying to run ssaha on a fedora 7 (32bit) system and it s dying with an error every time I run it. I m running the latest v33 version of ssaha, but I ve
    Message 1 of 5 , Aug 23 1:29 AM
    • 0 Attachment
      I'm trying to run ssaha on a fedora 7 (32bit) system and it's dying
      with an error every time I run it.

      I'm running the latest v33 version of ssaha, but I've also tried v31c
      which also fails.

      I can successfully compute the hashtable from my subject fasta files,
      but when I try to search against this with a short test sequence ssaha
      dies.

      /usr/local/bin/ssaha test.fa
      /data/public/DNA/Genomes/NCBIM36/Current/ncbi36_1 -qf fasta -sf hash

      [snip lots of output]

      Info: running DNA query against DNA hash table.
      ssaha:
      /home/andrewss/SSAHA/SSAHA/SequenceReader/SequenceReader.cpp:305:
      const char* SourceReaderIndex::extractName(SequenceNumber): Assertion
      `(nameBuffer_[0]=='>')||(nameBuffer_[0]=='@')' failed.
      Aborted

      Any ideas about how to make this work?
    • coxtonyj
      Hi Your command should work if test.fa is in valid fasta format. However the file needs to be in UNIX text file format - ie just one character at each
      Message 2 of 5 , Aug 23 4:28 AM
      • 0 Attachment
        Hi

        Your command should work if test.fa is in valid fasta format. However
        the file needs to be in UNIX text file format - ie just one character
        at each linebreak. I managed to replicate this error by giving it a
        fasta file in DOS text file format.

        Try 'dos2unix test.fa' and rerunning the query.

        Cheers


        Tony




        --- In ssahausers@yahoogroups.com, "sraxylan" <sraxylan@...> wrote:
        >
        > I'm trying to run ssaha on a fedora 7 (32bit) system and it's dying
        > with an error every time I run it.
        >
        > I'm running the latest v33 version of ssaha, but I've also tried
        v31c
        > which also fails.
        >
        > I can successfully compute the hashtable from my subject fasta
        files,
        > but when I try to search against this with a short test sequence
        ssaha
        > dies.
        >
        > /usr/local/bin/ssaha test.fa
        > /data/public/DNA/Genomes/NCBIM36/Current/ncbi36_1 -qf fasta -sf hash
        >
        > [snip lots of output]
        >
        > Info: running DNA query against DNA hash table.
        > ssaha:
        > /home/andrewss/SSAHA/SSAHA/SequenceReader/SequenceReader.cpp:305:
        > const char* SourceReaderIndex::extractName(SequenceNumber):
        Assertion
        > `(nameBuffer_[0]=='>')||(nameBuffer_[0]=='@')' failed.
        > Aborted
        >
        > Any ideas about how to make this work?
        >
      • sraxylan
        ... Tony, Thanks for the reply. I ve checked, and test.fa is definitely using unix line endings already. Having played with this a bit more it seems this is
        Message 3 of 5 , Aug 23 6:02 AM
        • 0 Attachment
          --- In ssahausers@yahoogroups.com, "coxtonyj" <anthony.j.cox@...> wrote:
          >
          > Hi
          >
          > Your command should work if test.fa is in valid fasta format.
          > However the file needs to be in UNIX text file format - ie
          > just one character at each linebreak. I managed to replicate
          > this error by giving it a fasta file in DOS text file format.
          >
          > Try 'dos2unix test.fa' and rerunning the query.

          Tony,

          Thanks for the reply. I've checked, and test.fa is definitely using
          unix line endings already.

          Having played with this a bit more it seems this is something to do
          with using absolute paths. If I run a search where the query and the
          hash are in the same folder then it works. If I leave the hash in one
          folder and move the query to another and then run a command which uses
          an absolute path to the hash then it fails. Am I missing something in
          my setup which would allow me to run searches from a directory which
          doesn't contain the search database?

          Cheers

          Simon.


          >
          > --- In ssahausers@yahoogroups.com, "sraxylan" <sraxylan@> wrote:
          > >
          > > I'm trying to run ssaha on a fedora 7 (32bit) system and it's dying
          > > with an error every time I run it.
          > >
          > > I'm running the latest v33 version of ssaha, but I've also tried
          > v31c
          > > which also fails.
          > >
          > > I can successfully compute the hashtable from my subject fasta
          > files,
          > > but when I try to search against this with a short test sequence
          > ssaha
          > > dies.
          > >
          > > /usr/local/bin/ssaha test.fa
          > > /data/public/DNA/Genomes/NCBIM36/Current/ncbi36_1 -qf fasta -sf hash
          > >
          > > [snip lots of output]
          > >
          > > Info: running DNA query against DNA hash table.
          > > ssaha:
          > > /home/andrewss/SSAHA/SSAHA/SequenceReader/SequenceReader.cpp:305:
          > > const char* SourceReaderIndex::extractName(SequenceNumber):
          > Assertion
          > > `(nameBuffer_[0]=='>')||(nameBuffer_[0]=='@')' failed.
          > > Aborted
          > >
          > > Any ideas about how to make this work?
          > >
          >
        • coxtonyj
          You re right, there does seem to be an issue when you specify -sf hash - both query and subject need to be in the same directory. Interestingly it always
          Message 4 of 5 , Aug 23 7:19 AM
          • 0 Attachment
            You're right, there does seem to be an issue when you specify '-sf
            hash' - both query and subject need to be in the same directory.
            Interestingly it always seems to be able to load in the hash table
            correctly, even if that is the one with the path. I will try to look
            into it.

            If both query and subject are '-sf fasta' then any combination of
            relative or absolute paths for both query and subject file works OK
            for me.

            --- In ssahausers@yahoogroups.com, "sraxylan" <simon.andrews@...>
            wrote:
            >
            > --- In ssahausers@yahoogroups.com, "coxtonyj" <anthony.j.cox@>
            wrote:
            > >
            > > Hi
            > >
            > > Your command should work if test.fa is in valid fasta format.
            > > However the file needs to be in UNIX text file format - ie
            > > just one character at each linebreak. I managed to replicate
            > > this error by giving it a fasta file in DOS text file format.
            > >
            > > Try 'dos2unix test.fa' and rerunning the query.
            >
            > Tony,
            >
            > Thanks for the reply. I've checked, and test.fa is definitely using
            > unix line endings already.
            >
            > Having played with this a bit more it seems this is something to do
            > with using absolute paths. If I run a search where the query and
            the
            > hash are in the same folder then it works. If I leave the hash in
            one
            > folder and move the query to another and then run a command which
            uses
            > an absolute path to the hash then it fails. Am I missing something
            in
            > my setup which would allow me to run searches from a directory which
            > doesn't contain the search database?
            >
            > Cheers
            >
            > Simon.
            >
            >
            > >
            > > --- In ssahausers@yahoogroups.com, "sraxylan" <sraxylan@> wrote:
            > > >
            > > > I'm trying to run ssaha on a fedora 7 (32bit) system and it's
            dying
            > > > with an error every time I run it.
            > > >
            > > > I'm running the latest v33 version of ssaha, but I've also
            tried
            > > v31c
            > > > which also fails.
            > > >
            > > > I can successfully compute the hashtable from my subject fasta
            > > files,
            > > > but when I try to search against this with a short test
            sequence
            > > ssaha
            > > > dies.
            > > >
            > > > /usr/local/bin/ssaha test.fa
            > > > /data/public/DNA/Genomes/NCBIM36/Current/ncbi36_1 -qf fasta -sf
            hash
            > > >
            > > > [snip lots of output]
            > > >
            > > > Info: running DNA query against DNA hash table.
            > > > ssaha:
            > >
            > /home/andrewss/SSAHA/SSAHA/SequenceReader/SequenceReader.cpp:305:
            > > > const char* SourceReaderIndex::extractName(SequenceNumber):
            > > Assertion
            > > > `(nameBuffer_[0]=='>')||(nameBuffer_[0]=='@')' failed.
            > > > Aborted
            > > >
            > > > Any ideas about how to make this work?
            > > >
            > >
            >
          • sraxylan
            ... Having looked at this some more I found that if you use a full path to specify the location of the fasta file when you create the hash initially you can
            Message 5 of 5 , Aug 24 3:14 AM
            • 0 Attachment
              --- In ssahausers@yahoogroups.com, "coxtonyj" <anthony.j.cox@...> wrote:
              >
              > You're right, there does seem to be an issue when you specify '-sf
              > hash' - both query and subject need to be in the same directory.
              > Interestingly it always seems to be able to load in the hash table
              > correctly, even if that is the one with the path. I will try to look
              > into it.

              Having looked at this some more I found that if you use a full path to
              specify the location of the fasta file when you create the hash
              initially you can then perform searches from anywhere. This gives me
              a good work round for my problem. It might be a simple fix to expand
              the filename to a full path when saving a hash?

              TTFN

              Simon.
            Your message has been successfully submitted and would be delivered to recipients shortly.