FEATURE
See COPYING and AUTHORS for copyright and license agreement terms.

----

See README.html for background and instructions.


FEATURE programs
----------------
featurize
Description:
    Generates microenvironments around sites.
Examples:
    featurize -P test.points > test.features
    featurize 1cll
    featurize -f pdbidlist
Usage: featurize [OPTIONS] [PDBID...]
Options:
    -v  Increase verbosity
    -n NUMSHELLS [Default: 6]
        Set number of shells to NUMSHELLS for environment mode
    -w SHELLWIDTH [Default: 1.25]
        Set thickness of each shell to SHELLWIDTH Angstroms
        for environment mode
    -x EXCLUDERESIDUES [Default: HETATM]
        Set residues to exclude to comma separated list of
        EXCLUDERESIDUES
    -f PDBIDFILE
        Read PDBIDs from PDBIDFILE
    -m FEATUREMODE [Default: environment]
        Set mode for feature extraction to FEATUREMODE
        Modes: atom environment labels
    -P POINTFILE
        Read point list from POINTFILE
    -H  Print header

buildmodel
Description:
    Creates score matrix for a given training set.
Examples:
    buildmodel -s 0.001 trainsite.features trainnonsite.features > train.model
    buildmodel site.features nonsite.features
Usage: buildmodel [OPTIONS] SITEFEATURES NONSITEFEATURES
Options:
    -v  Increase verbosity
    -n NUMSHELLS [Default: 6]
        Set number of shells to NUMSHELLS for environment mode
    -b NUMBINS [Default: 5]
        Set number of bins
    -s SITEPRIOR [Default: 0.01]
        Set the site prior

scoreit
Description:
    Scores feature vectors with a scoring model
Examples:
    scoreit -a train.model test.features > test.scores
Usage: scoreit [OPTIONS] SCOREFILE ENVIRONMENTFILE
Options:
    -v  Increase verbosity
    -n NUMSHELLS [Default: 6]
        Set number of shells
    -p PLEVEL [Default: 0.01]
        Set p-level of scorefile
    -c CUTOFF [Default: 0]
        Set minimum score to print out
    -a  Shows all scores
    -e  Count empty shells

scorefeatures

Deprecated FEATURE programs
---------------------------
training
scan
convert
prekin