Cleaning obsolete code/comments

In particular: Removed code for producing LateX-figures of thesis and
for assessing the contributions of the new start-, intron- and
stop-models (this code will be kept in my masterthesis-repository)

src/main/java/de/vetter/pogigwasc/App.java

@@ -19,18 +19,19 @@
 import java.time.format.DateTimeFormatter;
 
 /**
- * Main class for gene-prediction: Takes at least one, and up to three
+ * Main class for gene-prediction: Takes at least one, and up to four
  * parameters:
  * <ol>
  * <li>-i: the input-file (.fasta)</li>
  * <li>-o: the output file (.gff, can have arbitrary extension)</li>
  * <li>-p: the parameter-file (.properties), flagged by "-p"</li>
+ * <li>-n: flag whether the prediction should be intronless (this does
+ * <b>not</b> quite mean prediction optimized for mRNA-reads, as multiple genes
+ * can still be predicted on the same contig!)
  * </ol>
  * 
  * If no output file is specified, one will be created in the same directory as
  * the input-file, with date and time-information in its filename.<br>
- * If no parameter-file is specified, a default parameter-file (also, and mainly
- * used for code-testing) will be used (<b>this no longer works, if the executable is removed from the sources!</b>)
  */
 public class App {
 

src/main/java/de/vetter/pogigwasc/LoxodesMagnusGHMM.java

@@ -182,12 +182,11 @@ public static String parseToGFF(String header, Parse parse, ModelParameters para
 					result.append(System.lineSeparator());
 
 				}
-
+				// Handling the start-region which extends beyond the CDS
 				if (currentFeature != GFFFeature.CDS)
 					startOfCurrentFeature = state.getName().equals(LoxodesMagnusGHMM.FORWARD_START)
 							? currentPos + (parameters.getStartRegionSize() - 3)
 							: currentPos;
-				// TODO ^ notice this! It is for the new StartRegion-state!
 				currentFeature = GFFFeature.CDS;
 
 				break;

src/main/java/de/vetter/pogigwasc/ModelParameters.java

@@ -62,8 +62,6 @@ public ModelParameters(FileReader parameterFileReader) throws IOException {
 
 			currentParameter = "start_region_size";
 			startregionSize = Integer.parseInt(properties.getProperty(currentParameter));
-			// old start-region (i.e. just deterministic ATG):
-			// startregionSize = 3;
 
 			currentParameter = "stop_region_size";
 			stopregionSize = Integer.parseInt(properties.getProperty(currentParameter));
@@ -310,22 +308,6 @@ private void computePoissonNormaliser() {
 		intronTruncatedPoissonNormaliser = 1-Math.exp(intronTruncatedPoissonNormaliser);
 		intronTruncatedPoissonNormaliser = intronTruncatedPoissonNormaliser/(intronMax - intronMin + 1);
 		// This value will be ADDED to the probability of seeing any length within min-max.
-		/*
-		System.out.println("Intron-state: length-distribution:");
-		System.out.println("l,poisson,empirical");
-
-		double[] LENGTH_PROBABILITIES = new double[] { 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,
-				0.0, 0.0, 0.0, 0.0, 0.2188, 0.2812, 0.1042, 0.2292, 0.1146, 0.0208, 0.0104,
-				0.0104, 0.0, 0.0, 0.0104, 0.0, 0.0, 0.0 };
-		
-		double sum = 0;
-		for(int k = 0; k < intronMax + 2; k++) {
-			sum += Math.exp(getLogProbabilityIntronLength(k));
-			System.out.println(k + "," + Math.exp(getLogProbabilityIntronLength(k)) + ","
-					+ (k < LENGTH_PROBABILITIES.length ? LENGTH_PROBABILITIES[k] : 0));
-		}
-		System.out.println("\n sum = " + sum + " (should be 1)");
-		*/
 	}
 
 	/**

src/main/java/de/vetter/pogigwasc/Utilities.java

@@ -9,7 +9,7 @@
  */
 public class Utilities {
 
-	// Can be used for poisson -- but currently, i am no longer using poisson.
+	// Used for Poisson
 	private static final double[] LOG_FACTORIALS = new double[] { 0.0, 0.0, 0.6931471805599453, 1.791759469228055,
 			3.1780538303479453, 4.787491742782046, 6.579251212010101, 8.525161361065415, 10.60460290274525,
 			12.80182748008147, 15.104412573075518, 17.502307845873887, 19.98721449566189, 22.552163853123425,

src/main/java/de/vetter/pogigwasc/Viterbi.java

@@ -3,12 +3,10 @@
 import java.util.ArrayList;
 import java.util.List;
 
-import de.vetter.pogigwasc.states.HMMState;
-
 /**
- * Implements the viterbi algorithm as described by Stanke (AUGUSTUS): Construct
+ * Implements the Viterbi algorithm as described by Stanke (AUGUSTUS): Construct
  * a Viterbi-instance for a GHMM and a given sequence, to compute the
- * viterbi-variables of that model on that sequence, and yield the MAP-parses
+ * Viterbi-variables of that model on that sequence, and yield the MAP-parses
  * 
  * @author David Emanuel Vetter
  *

src/main/java/de/vetter/pogigwasc/states/IntronState.java

@@ -64,12 +64,6 @@ public double computeLogEmissionProbability(int previousState, String emissionHi
 			// if(baseUsage == Double.NEGATIVE_INFINITY)
 			// 	return baseUsage;
 		} 
-		/* Compute without respect to intron-structure! This is the old style 
-		baseUsage = 0;
-		for(int i = 2; i < length - 2; i++) {
-			baseUsage += parameters.getLogBaseProbabilityNCS(newEmission.charAt(i));
-		}
-		*/
 
 		return lengthProb + baseUsage;
 	}

src/main/java/de/vetter/pogigwasc/states/StartRegionState.java

@@ -26,7 +26,6 @@ public double computeLogEmissionProbability(int previousState, String emissionHi
 		}
 
 		double result = 0;
-		/* */
 		for(int i = 0; i < parameters.getStartRegionSize() - 3; i++) {
 			result += parameters.getLogBaseProbabilityStartRegion(newEmission.charAt(i));
 		}

src/main/java/de/vetter/pogigwasc/states/StopRegionState.java

@@ -27,32 +27,6 @@ public double computeLogEmissionProbability(int previousState, String emissionHi
 		for(int i = 0; i < parameters.getStopRegionSize() - 3; i += 3)
 			result += parameters.getLogCodonProbabilityStopRegion(newEmission.substring(i, i+3));
 
-		/* old mode:  
-		double[] empirical = { 0.03837953091684435, 0.029850746268656716, 0.0255863539445629, 0.019189765458422176,
-				0.014925373134328358, 0.008528784648187633, 0.010660980810234541, 0.0021321961620469083,
-				0.029850746268656716, 0.0255863539445629, 0.0042643923240938165, 0.017057569296375266,
-				0.010660980810234541, 0.008528784648187633, 0.0, 0.01279317697228145, 0.023454157782515993,
-				0.014925373134328358, 0.021321961620469083, 0.0042643923240938165, 0.029850746268656716,
-				0.01279317697228145, 0.023454157782515993, 0.006396588486140725, 0.008528784648187633,
-				0.006396588486140725, 0.019189765458422176, 0.0042643923240938165, 0.008528784648187633,
-				0.0021321961620469083, 0.0021321961620469083, 0.0021321961620469083, 0.0255863539445629,
-				0.014925373134328358, 0.006396588486140725, 0.021321961620469083, 0.017057569296375266,
-				0.006396588486140725, 0.008528784648187633, 0.0042643923240938165, 0.0255863539445629,
-				0.017057569296375266, 0.05543710021321962, 0.042643923240938165, 0.008528784648187633,
-				0.006396588486140725, 0.042643923240938165, 0.008528784648187633, 0.02771855010660981,
-				0.01279317697228145, 0.006396588486140725, 0.006396588486140725, 0.031982942430703626,
-				0.0021321961620469083, 0.021321961620469083, 0.006396588486140725, 0.019189765458422176,
-				0.019189765458422176, 0.02771855010660981, 0.017057569296375266, 0.008528784648187633,
-				0.006396588486140725, 0.021321961620469083, 0.006396588486140725 };
-		result = 0;
-		for(int i = 0; i < parameters.getStopRegionSize() - 3; i += 3) {
-			int codonIndex = Utilities.baseToIndex(newEmission.charAt(i)) * 16;
-			codonIndex += Utilities.baseToIndex(newEmission.charAt(i+1)) * 4;
-			codonIndex += Utilities.baseToIndex(newEmission.charAt(i+2));
-			result += empirical[codonIndex];
-		}
-		// END old mode */
-
 		return result;
 	}