001 package calhoun.analysis.crf.features.interval29;
002
003 import calhoun.analysis.crf.CacheStrategySpec;
004 import calhoun.analysis.crf.FeatureManagerNodeBoundaries;
005 import calhoun.analysis.crf.CacheStrategySpec.CacheStrategy;
006
007
008 public class ReferenceBasePredictorInterval29 extends ReferenceBasePredictorInterval29Base implements FeatureManagerNodeBoundaries<Character> {
009
010 private static final long serialVersionUID = -8460452348450096338L;
011
012 public ReferenceBasePredictorInterval29() {
013 }
014
015 public CacheStrategySpec getCacheStrategy() {
016
017 CacheStrategySpec css = new CacheStrategySpec(CacheStrategy.DENSE_NODE_BOUNDARY);
018 CacheStrategySpec.DenseBoundaryCachingDetails details = new CacheStrategySpec.DenseBoundaryCachingDetails(25); // we will use 9 tables...nix. 25 for interval29
019
020
021 // If you want to predict for real and mesh with a PWM, then set pads, for example, as below:
022 // donor[j] = new PWMLookup(3,6,pseudoCounts); // donor signal xxx|GTxxxx
023 // acceptor[j] = new PWMLookup(9,6,pseudoCounts); // acceptor signal xxxxxxxAG|xxxxxx
024 // start = new PWMLookup(9,6,pseudoCounts); // start signal xxxxxxxxx|ATGxxx
025 // stop = new PWMLookup(3,9,pseudoCounts); // stop signal xxx|TAGxxxxxx
026
027 // IDEALLY HAVE SOME ASSERTIONS TYING THESE PADS TO THE PWM SIZES
028 details.add(0,0,startIx + (multipleFeatures ? 0 : 0),Interval29Model.getPadIntergenic(),Interval29Model.getPadIntergenic()); // minlength 18
029 details.add(1,1,startIx + (multipleFeatures ? 1 : 0),Interval29Model.getPadExon3prime(),Interval29Model.getPadExon5prime()); // minlength 9
030 details.add(2,2,startIx + (multipleFeatures ? 1 : 0),Interval29Model.getPadExon3prime(),Interval29Model.getPadExon5prime());
031 details.add(3,3,startIx + (multipleFeatures ? 1 : 0),Interval29Model.getPadExon3prime(),Interval29Model.getPadExon5prime());
032 details.add(4,4,startIx + (multipleFeatures ? 2 : 0),Interval29Model.getPadIntron3prime(),Interval29Model.getPadIntron5prime()); // minlength 15
033 details.add(5,4,startIx + (multipleFeatures ? 2 : 0),Interval29Model.getPadIntron3prime(),Interval29Model.getPadIntron5prime());
034 details.add(6,4,startIx + (multipleFeatures ? 2 : 0),Interval29Model.getPadIntron3prime(),Interval29Model.getPadIntron5prime());
035 details.add(7,5,startIx + (multipleFeatures ? 3 : 0),Interval29Model.getPadExon5prime(),Interval29Model.getPadExon3prime()); // minlength 9
036 details.add(8,6,startIx + (multipleFeatures ? 3 : 0),Interval29Model.getPadExon5prime(),Interval29Model.getPadExon3prime());
037 details.add(9,7,startIx + (multipleFeatures ? 3 : 0),Interval29Model.getPadExon5prime(),Interval29Model.getPadExon3prime());
038 details.add(10,8,startIx + (multipleFeatures ? 4 : 0),Interval29Model.getPadIntron5prime(),Interval29Model.getPadIntron3prime()); // minlength 15
039 details.add(11,8,startIx + (multipleFeatures ? 4 : 0),Interval29Model.getPadIntron5prime(),Interval29Model.getPadIntron3prime());
040 details.add(12,8,startIx + (multipleFeatures ? 4 : 0),Interval29Model.getPadIntron5prime(),Interval29Model.getPadIntron3prime());
041 details.add(13,0,startIx + (multipleFeatures ? 0 : 0),0,0);
042 details.add(14,0,startIx + (multipleFeatures ? 0 : 0),0,0);
043 details.add(15,4,startIx + (multipleFeatures ? 2 : 0),0,0);
044 details.add(16,4,startIx + (multipleFeatures ? 2 : 0),0,0);
045 details.add(17,4,startIx + (multipleFeatures ? 2 : 0),0,0);
046 details.add(18,4,startIx + (multipleFeatures ? 2 : 0),0,0);
047 details.add(19,4,startIx + (multipleFeatures ? 2 : 0),0,0);
048 details.add(20,4,startIx + (multipleFeatures ? 2 : 0),0,0);
049 details.add(21,0,startIx + (multipleFeatures ? 0 : 0),0,0);
050 details.add(22,0,startIx + (multipleFeatures ? 0 : 0),0,0);
051 details.add(23,8,startIx + (multipleFeatures ? 4 : 0),0,0);
052 details.add(24,8,startIx + (multipleFeatures ? 4 : 0),0,0);
053 details.add(25,8,startIx + (multipleFeatures ? 4 : 0),0,0);
054 details.add(26,8,startIx + (multipleFeatures ? 4 : 0),0,0);
055 details.add(27,8,startIx + (multipleFeatures ? 4 : 0),0,0);
056 details.add(28,8,startIx + (multipleFeatures ? 4 : 0),0,0);
057 // NOTE: The minimum lengths of the Semi-Markov training and inference for each state must be AT least as big as the sum of the two pads at either end.
058
059 details.check();
060
061 css.details = details;
062
063 return css;
064 }
065 }