This feature is not fully developed, but it is useable (and useful!). For pairwise comparisons, the .seq file MUST NOT contain any non-amino acid characters or spaces in the sequence part of the file.
Having checked this, you must first create a copy of the .seq file (call this .sec). The .sec file could contain secondary structure definitions for the protein, or any other characters that you want to align with the sequences. Check that your SCANPS defaults file has the value of MAX_NSEQ set greater than the number of sequences in your sequence file, then for example, for the file test.seq type:
scanps -stest.seq -ttest.sec -T
This gives the score for each pair comparison to stdout. You could redirect the output to a file.
553 HAJUA HAHOD 543 HAJUA HAHOK 475 HAJUA HAKOAW 481 HAJUA HAJSA 461 HAJUA HAFEDR 261 HAJUA HBOTE 646 HAHOD HAHOK 490 HAHOD HAKOAW 502 HAHOD HAJSA 471 HAHOD HAFEDR 306 HAHOD HBOTE 484 HAHOK HAKOAW 490 HAHOK HAJSA 461 HAHOK HAFEDR 292 HAHOK HBOTE 587 HAKOAW HAJSA 433 HAKOAW HAFEDR 269 HAKOAW HBOTE 439 HAJSA HAFEDR 274 HAJSA HBOTE 307 HAFEDR HBOTE
Each line of the output shows the score and a the corresponding pair of ID codes.
Pairwise comparisons may also be performed using the NALL method. Currently, this only works if you also request probability scores. For example:
scanps -stest.seq -ttest.sec -T -a1 -F1
gives ...
7.2765e-88 HAJUA HAHOD 1.0265e-85 HAJUA HAHOK 2.5983e-71 HAJUA HAKOAW 1.4448e-72 HAJUA HAJSA 2.1362e-68 HAJUA HAFEDR 2.0553e-29 HAJUA HBOTE 3.4899e-108 HAHOD HAHOK 1.868e-74 HAHOD HAKOAW 5.5253e-77 HAHOD HAJSA 1.7759e-70 HAHOD HAFEDR 1.2143e-37 HAHOD HBOTE 3.3974e-73 HAHOK HAKOAW 1.868e-74 HAHOK HAJSA 2.1362e-68 HAHOK HAFEDR 4.8684e-35 HAHOK HBOTE 3.1609e-95 HAKOAW HAJSA 1.2637e-62 HAKOAW HAFEDR 7.5967e-31 HAKOAW HBOTE 7.4395e-64 HAJSA HAFEDR 9.5138e-32 HAJSA HBOTE 7.8891e-38 HAFEDR HBOTE
You can also get the alignments corresponding to these pair comparisons by adding the -v command line argument.
scanps -stest.seq -ttest.sec -T -a1 -F1 -v
The output of this comparison will include the characters from the .sec file aligned along with the sequences.
The final option in pairwise mode is to output the scores in a form that can be analysed by the cluster analysis program ``oc''. To produce suitable output, simply add a -X to the command line.
For example:
scanps -stest.seq -ttest.sec -T -X for raw scores or: scanps -stest.seq -ttest.sec -T -a1 -E -F1 -X for probabilities
The -E option is necessary to prevent scanps from writing all local alignment scores. For cluster analysis you only need the top scoring alignment.