Compile:
cc -O3 eigencontact_fread_threader.c -lm -o eigenthreader
cc -O3 strsum_eigen.c -lm -o strsum_eigen

Typical command line (similar to GenTHREADER):
eigenthreader -c12 -C0 -t20 -z1250 -m$target -Fss2files/$target.ss2 seqfiles/$target.fasta confiles/$target.con $target.out scop.lst

-c12 = contact distance 12A (something to benchmark perhaps?)
-t20 = 20 eigenvectors considered (something to benchmark in the paper i.e. try from 1 to n eigenvectors)
-C0 = minimum contact alignment score required to generate alignment/model
-z1250 = scaling of eigenvectors (again something to play with but I think this is a reasonable value)

$target.ss2 is standard PSIPRED output file
$target.con is MetaPSICOV output file (CASP RR format) (see confiles.tar.gz for PSICOV150 benchmark files)

-m will produce PDB file output based on alignments (as with GenTHREADER)
-p will output alignment on stdout

scop.lst is the fold library list as with GenTHREADER etc.

Also see example.csh.


Example to generate paired TDB and eigen file (outputs):

strsum_eigen 1jbeA.pdb 1jbeA.dssp $TDB_DIR/1jbeA.tdb $TDB_DIR/1jbeA.eig

(Env variable TDB_DIR specifies directory for fold library)

Output:

eigenthreader output file format. Columns are

1) Estimated probability
2) T-score
3) Alignment length
4) Percentage coverage query
5) Percentage coverage hit
6) Match name