configure

#this file has all the parameter which would not be assigned from command line for all the functions
#If you don't know the meaning of every one, while want to know weather a certain parameter important for the function you are running.
#You could just remove this parameter and run the interesting function. If the removed function does mater for the interesting function, the software would give waring message and stop.
# parameters for gene structure realignment scoring begin

alignmentOpenGapP   -4
alignmentExtendGapP -2

alignmentExonMatchP 6
alignmentExonMismatchP  -9
alignmentExonOpenGapP   -4
alignmentExonExtendGapP -2
alignmentIntronMatchP   1
alignmentIntronMismatchP    -1
alignmentIntronOpenGapP -2
alignmentIntronExtendGapP   -1
alignmentStartStopCodonMatchP   10
alignmentStartStopCodonMismatchP    -10
alignmentStartStopCodonOpenGapP -10
alignmentStartStopCodonExtendGapP   -10
alignmentSpliceSitesMatchP  0
alignmentSpliceSitesMismatchP   0
alignmentSpliceSitesOpenGapP    -10
alignmentSpliceSitesExtendGapP  -10
# parameters for gene structure realignment scoring end

# parameters for MSA scoring begin
# The MSA function is not implemented yet
# myMsaMatchP 5
# myMsaMisMatchP  -3
# myMsaGapP   -1
# myMsaOpenGapP   -3
# parameters for MSA scoring end

# folder for temporary files
tempFolder   ./temp
#regular to parse reference gff file. Get the map from transcript id to gene id
transcript_to_gene_regex_reference_gff  [\s\S]*?CDS[\s\S]*?Parent=((\S*?)\.\d+)
#regular to parse additional gff file. Get the map from transcript id to gene id
transcript_to_gene_regex_novo_gff   transcript[\s\S]*?ID=(\S*?);Parent=(\S+)
#regular to parse reference gff file. Get the map from CDS records to transcript id
#it seems C++ have problem with \w and \d
cdsParentRegex  ([\s\S]*)Parent=([abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789.:_-]+)
#regular to parse additional gff file. Get the map from CDS records to transcript id
novo_cdsParentRegex [\s\S]*?ID=([\s\S]*?);Parent=([\s\S]*?)$
# regular expression to get the file name of gff file. The file name would be used as the prefix for temp files
temp_file_regex  [\s\S]*[(\\)(\/)]([\s\S]*)
maxintronlength   30000
minintronlength   10
MsaPreFolder    ./MsaPreFolder