I am trying to build my own database with HHsearch. I have an MSA for each of the proteins I want to include in the database. When I try to add the secondary structure to each of the MSA by
perl /home/chris/Bioinformatics/hhsuite-2.0.16/scripts/multithread.pl '/home/chris/Bioinformatics/hhsuite-2.0.16/databases/VV/*.a3m' '/home/chris/Bioinformatics/hhsuite-2.0.16/scripts/addss.pl $file'
The "addss" script returns an error when it tries to predict the secondary structure of any of the proteins. The error I get is
Error: command '/cluster/toolkit/production/bioprogs/blast/bin/blastpgp -b 1 -j 1 -h 0.001 -d /home/chris/Bioinformatics/hhsuite-2.0.16/lib/hh/data/do_not_delete -i /tmp/CnHsQlKE_r/RPWl2LU3oC.sq -B /tmp/CnHsQlKE_r/RPWl2LU3oC.in.psi -C /tmp/CnHsQlKE_r/RPWl2LU3oC.chk 1> /tmp/CnHsQlKE_r/RPWl2LU3oC.blalog 2> /tmp/CnHsQlKE_r/RPWl2LU3oC.blalog' returned error code 127 $ echo RPWl2LU3oC.chk > /tmp/CnHsQlKE_r/RPWl2LU3oC.pn $ echo RPWl2LU3oC.chk > /tmp/CnHsQlKE_r/RPWl2LU3oC.pn $ /cluster/toolkit/production/bioprogs/blast/bin/makemat -P /tmp/CnHsQlKE_r/RPWl2LU3oC Error: failed to execute '/cluster/toolkit/production/bioprogs/blast/bin/makemat -P /tmp/CnHsQlKE_r/RPWl2LU3oC': No such file or directory
Could this be due to the fact, that I have psiblast but blastpgp installed?
Judging from the source code of the addss.pl script, calls psipred for prediciting the secondary structure. Within the PSIPRED scripts (runpsipred, runpsipredsingle and runpsipredplus) I changed the variables "execdir", "ncbidir", "dbname" and "datadir" as recommended in the PSIPRED installation guide. Do I have to set those variables differently if I call PSIPRED from a HHsearch script? Running PSIPRED without calling it from wihtin HHsearch works just fine. Concerning the missing file the error is pointing at: Is there any directory I have to specify either in the "addss" script or the PSIPRED executable?
I took a closer look at what happens. The addss.pl generates temporary files from the a3m-files of the protein sequences, just as it should. Right after starting the secondary structure predition , when psiblast should be applied to the sequences the script crashes. To me it looks like it is searching for PSIBLAT/BLASTPGP in a directory where PSIBLAST is not located, at least not on my system.
Does anyone have some advice on resolving this error?
Thank you a lot for your help!