|
||
| Home | People | Research | Publications | News |
|
Running Prospect We suggest users to run BLAST / PSI-BLAST
first before using PROSPECT to make sure whether any homolog for
the target exists in PDB. In case a remote homolog found by PSI-BLAST
has only alignment for partial sequence and it is not included in the
DALI or FSSP list, it is suggested to include them in the template library
(see Templates) to verify if it is the
true fold or to generate the full alignment. One can do the same thing
for a PDB structure having similar function to the target. The programs included as part of the prospect suite are:
For the various examples below, we use 'LINUX' as the architecture in the commands. Sequence ProfileWe use PSI-Blast to generate sequence profiles. The program
blastpgp produces a 'checkpoint file' at the end of a search iteration.
We include the script get_chk_file, which can be used with the command: <seq file> : The path to the sequence file This is produce the file <seq file>.chk You can run blastpgp yourself, with the command: blastpgp -b 0 -j 3 -h 0.001 -d /data/nr/nr -i test.seq -C test.seq.chk This tells blastpgp to show none of the alignments (-b 0), with
3 iterations (-j 3), and an e-value threshold of 0.001 (-h 0.001), using
a database found at /data/nr/nr (-d), the sequence test.seq (-i), and saving
the checkpoint file (thing we're intersted in here) to file test.seq.chk
(-C) The chk file is a non-ascii file which is not transferable between
machines with different encoding schemes (i.e. between big and little endian
machines). read_chk.<Architecture> <chk file> This will read all of the information in the file out to ASCII format.
To save it, simple pipe the output to a file, like This file should be comprised of one line with the number of amino
acids in the sequnce (N), followed by the sequnce on the next line, and
then N lines Secondary Structure Predictionprospect_ssp.<Architecture> [-freqfile <file>]/[-seqfile <file>]/[-chkfile <file>] [-p]-chkfile <file> The frequency profile for a sequence, created by psi-blast.-freqfile <file> The ASCII version of a check point (chk) file, created with the tool read_chk. -seqfile <file> The raw sequence, typically in standard FASTA format
-p Print output in PHD style format
AlignmentThreading is the actual 'work horse' of the prospect suite. It
is the program that actually does the actual threading procedure. Typically
you don't call threading directly, but rather call it through prospect, which
then calls threading aginst an entire database. -phdfile <file> A secondary structure prediction in PHD format, can be generated by prospect_ssp -seqfile <file> The amino acid sequence file, typically in FASTA format. Note, if you provide a secondary structure prediction, you don't need to provide this file.
-freqfile <file> The frequency profile for the sequence, as outputed by read_chk
-global -global_local -np -wp Select the type of threading procedure to use. See threading methods for more information.
-reliab Calculate the zscore when doing global and global local threading. Does not apply to NP and WP threading.
-o <output file> Name of the file to output to. [TemplateName ]/[-tempfile <file>]
e.g., The options at the command line are complementary with the settings specified in configuations.
Threading against a Database For fold recognition:
All of these parameters are passed directly to threading -scop Thread aginst the SCOP database.-fssp Thread aginst the FSSP database (this is the default action).
-custom Thread aginst all templates found in the template paths that are not part of FSSP or SCOP (i.e. templates that you've generated with make_template) -all Thread aginst all found templates.
-tfile <file>
-o <output file> Name of the output XML file-ncpus N Try to launch N simultanous threading jobs. e.g., Run default template list aginst sequence The following optional flags are available:
After ProspectSorting prospect ResultsSortProspect is provided to scan and sort the names of templates in prospect files in order to determine which templates are the best matches. sortProspect.<Architecture> <prospect outfile> [-r/-z] [-s] [-1] [-top x] <prospect outfile> The file outputed by prospect
-r Sort by raw score
-z Sort by zscore
-s Save the prospect output file with the enteries in the order of sorting
-1 One colum view, only print out the names of templates
-top x Print out the top x scores of the sort There are different methods to sort scores, amoung them: by raw score,
Zscore, and SVM score. SVM sort is default. Zscore, and raw score
can be activated by the -z and -r flags respectively. To sort according to the SVM score, and save sort back to the file:
Confidence Index
|
|||||||||||||||||
|
| Life Sciences Division - ORNL - Disclaimer - Webmaster |