"Fossies" - the Fresh Open Source Software Archive

Member "augustus-3.3.3/config/species/galdieria/galdieria_parameters.cfg" (22 May 2019, 6195 Bytes) of package /linux/misc/augustus-3.3.3.tar.gz:


As a special service "Fossies" has tried to format the requested text file into HTML format (style: standard) with prefixed line numbers. Alternatively you can here view or download the uninterpreted source code file. See also the latest Fossies "Diffs" side-by-side code changes report for "galdieria_parameters.cfg": 3.3.2_vs_3.3.3.

    1 #
    2 # parameters for Galdieria sulphuraria (red microalga)
    3 # 
    4 # date : 24.06.2005
    5 #
    6 
    7 #
    8 # Properties for augustus
    9 #------------------------------------
   10 /augustus/verbosity 3     # 0-3, 0: only print the necessary
   11 
   12 stopCodonExcludedFromCDS false # make this 'true' if the CDS in the training set do not include the stop codon
   13 
   14 # gff output options:
   15 protein             on    # output predicted protein sequence
   16 codingseq           on    # output the coding sequence
   17 exonnames           on    # output 'single', 'initial', 'terminal' or 'internal' for CDS
   18 cds                 on    # output 'cds' as feature for exons
   19 start               on    # output start codons (translation start)
   20 stop                on    # output stop codons  (translation stop)
   21 introns             on    # output introns
   22 print_utr           on    # output 5'UTR and 3'UTR lines instead of exon lines
   23 
   24 checkExAcc          off   # internal parameter for extrinsic accuracy
   25 
   26 # alternative transcripts and posterior probabilities
   27 sample                      100   # the number of sampling iterations
   28 alternatives-from-sampling false  # output suboptimal alternative transcripts based on sampling
   29 alternatives-from-evidence true   # output alternative transcripts based on conflicting hints
   30 minexonintronprob           0.2   # minimal posterior probability of all (coding) exons
   31 minmeanexonintronprob       0.5   # minimal geometric mean of the posterior probs of introns and exons
   32 maxtracks                   2     # maximum number of reported transcripts per gene (-1: no limit)
   33 keep_viterbi                true  # set to true if all Viterbi transcripts should be reported
   34 
   35 UTR                         on
   36 
   37 #
   38 # Don't mess with any parameters below this point!
   39 # They are mainly used for training.
   40 
   41 # global constants
   42 # ----------------------------
   43 
   44 /Constant/trans_init_window           0
   45 /Constant/ass_upwindow_size           15
   46 /Constant/ass_start                   1
   47 /Constant/ass_end                     1
   48 /Constant/dss_start                   1
   49 /Constant/dss_end                     4
   50 /Constant/init_coding_len	      15
   51 /Constant/tss_upwindow_size           0
   52 /Constant/intterm_coding_len	      0
   53 /Constant/decomp_num_at               1
   54 /Constant/decomp_num_gc               1
   55 /Constant/gc_range_min  	      0.32
   56 /Constant/gc_range_max		      0.73
   57 /Constant/decomp_num_steps            1
   58 /Constant/min_coding_len              201
   59 /Constant/probNinCoding               0.24
   60 /Constant/amberprob                   .23      # Prob(stop codon = tag), if 0 tag is assumed to code for amino acid
   61 /Constant/ochreprob                   .46      # Prob(stop codon = taa), if 0 taa is assumed to code for amino acid
   62 /Constant/opalprob                    .31      # Prob(stop codon = tga), if 0 tga is assumed to code for amino acid
   63 /Constant/subopt_transcript_threshold 0.86
   64 /Constant/almost_identical_maxdiff    50
   65 /Constant/max_contra_supp_ratio       5.0      # lower than standard -> predict fewer alternatives
   66 
   67 # type of weighing, one of  1 = equalWeights, 2 = gcContentClasses, 3 = multiNormalKernel
   68 /BaseCount/weighingType        3
   69 # file with the weight matrix (only for multiNormalKernel type weighing)
   70 /BaseCount/weightMatrixFile   galdieria_weightmatrix.txt
   71 
   72 # Properties for IGenicModel
   73 # ----------------------------
   74 /IGenicModel/verbosity      3
   75 /IGenicModel/infile         galdieria_igenic_probs.pbl
   76 /IGenicModel/outfile        galdieria_igenic_probs.pbl.wech
   77 /IGenicModel/patpseudocount 4.5
   78 /IGenicModel/k              4
   79 #/IGenicModel/seqfile        ir.fa
   80 
   81 # Properties for ExonModel
   82 # ----------------------------
   83 /ExonModel/verbosity          3
   84 /ExonModel/infile             galdieria_exon_probs.pbl
   85 /ExonModel/outfile            galdieria_exon_probs.pbl.wech
   86 /ExonModel/patpseudocount     5
   87 /ExonModel/minPatSum          433
   88 /ExonModel/k                  4
   89 /ExonModel/etorder	      3
   90 /ExonModel/etpseudocount      3
   91 /ExonModel/exonlengthD        3000
   92 /ExonModel/maxexonlength      15000
   93 /ExonModel/slope_of_bandwidth 0.1875
   94 /ExonModel/minwindowcount     12
   95 /ExonModel/tis_motif_memory   2
   96 /ExonModel/tis_motif_radius   0
   97  
   98 # Properties for IntronModel
   99 # ----------------------------
  100 /IntronModel/verbosity          3
  101 /IntronModel/infile             galdieria_intron_probs.pbl
  102 /IntronModel/outfile            galdieria_intron_probs.pbl.wech
  103 /IntronModel/patpseudocount     1
  104 /IntronModel/k                  4
  105 /IntronModel/slope_of_bandwidth 0.1
  106 /IntronModel/minwindowcount     2
  107 /IntronModel/asspseudocount     0.0016
  108 /IntronModel/dsspseudocount     0.0001
  109 /IntronModel/dssneighborfactor  0.001325
  110 #/IntronModel/splicefile         galdieria_splicefile
  111 /IntronModel/sf_with_motif      true
  112 /IntronModel/d                  90
  113 /IntronModel/ass_motif_memory   2
  114 /IntronModel/ass_motif_radius   3
  115 
  116 # Properties for UtrModel
  117 # ----------------------------
  118 /UtrModel/verbosity             3
  119 /UtrModel/infile                galdieria_utr_probs.pbl
  120 /UtrModel/outfile               galdieria_utr_probs.pbl.wech
  121 /UtrModel/k                     3
  122 /UtrModel/utr5patternweight     1
  123 /UtrModel/utr3patternweight     0.5
  124 /UtrModel/patpseudocount        1
  125 /UtrModel/tssup_k               0
  126 /UtrModel/tssup_patpseudocount  1
  127 /UtrModel/slope_of_bandwidth    0.225
  128 /UtrModel/minwindowcount        5
  129 /UtrModel/exonlengthD           275
  130 /UtrModel/maxexonlength         700
  131 /UtrModel/max3singlelength      700
  132 /UtrModel/max3termlength        700
  133 /UtrModel/tss_start             0
  134 /UtrModel/tss_end               0
  135 /UtrModel/tata_start            0     # bases before tata
  136 /UtrModel/tata_end              0     # bases starting with first base of tata
  137 /UtrModel/d_tss_tata_min        0     # minimal distance between start of tata box (if existent) and tss 
  138 /UtrModel/d_tss_tata_max        0     # maximal distance between start of tata box (if existent) and tss
  139 /UtrModel/d_polyasig_cleavage   15      # the transcription end is predicted this many bases after the polyadenylation signal
  140 /UtrModel/d_polya_cleavage_min  10      # In training polyA singals are searched in this
  141 /UtrModel/d_polya_cleavage_max  25      # distance range range from the tts (cleavage site).
  142 /UtrModel/prob_polya            0.5
  143 /UtrModel/tts_motif_memory      1