/**
* Sample and calculate the probability of hitting each type of marker (marker.class). Creates
* 'numReads' reads of size 'readLen' and count how many of them hit each marker type.
*/
CountByType randomSampling(int readLen, int numReads) {
CountByType countReads = new CountByType();
RandMarker randMarker = new RandMarker(snpEffectPredictor.getGenome());
for (int i = 0; i < numReads; i++) {
// Random read
Marker read = randMarker.rand(readLen);
// Where does it hit?
Markers regions = snpEffectPredictor.queryDeep(read);
HashSet<String> doneRegion = new HashSet<String>();
for (Marker m : regions) {
String mtype = markerTypes.getType(m);
String msubtype = markerTypes.getSubType(m);
if (!doneRegion.contains(mtype)) {
countReads.inc(mtype); // Count reads
doneRegion.add(mtype); // Do not count twice
}
if ((msubtype != null) && !doneRegion.contains(msubtype)) {
countReads.inc(msubtype); // Count reads
doneRegion.add(msubtype); // Do not count twice
}
}
}
return countReads;
}
/** Build a genome from a GFF3 file and compare results to 'expected' results */
public SnpEffectPredictor buildAndCompare(
String genome,
String gff3File,
String resultFile,
boolean readSeqs,
boolean createRandSequences) {
String expectedResult = (resultFile == null ? "" : Gpr.readFile(resultFile).trim());
// Build
Config config = new Config(genome, Config.DEFAULT_CONFIG_FILE);
SnpEffPredictorFactoryGff3 fgff3 = new SnpEffPredictorFactoryGff3(config);
fgff3.setVerbose(verbose);
fgff3.setFileName(gff3File);
fgff3.setReadSequences(readSeqs);
fgff3.setCreateRandSequences(createRandSequences);
fgff3.setRandom(
new Random(
20140410)); // Note: we want consistent results in our test cases, so we always
// initialize the random generator in the same way
SnpEffectPredictor sep = fgff3.create();
// Compare result
String result = show(sep.getGenome()).trim();
if (verbose || !Gpr.noSpaces(expectedResult).equals(Gpr.noSpaces(result)))
System.out.println("Result:\n----------\n" + result + "\n----------\n");
Assert.assertEquals(Gpr.noSpaces(expectedResult), Gpr.noSpaces(result));
return sep;
}
/** Count bases covered for each marker type */
public void countBases() {
// ---
// Add all markers
// ---
Markers markers = new Markers();
markers.add(snpEffectPredictor.getMarkers());
for (Gene gene : snpEffectPredictor.getGenome().getGenes()) {
markers.add(gene);
markers.add(gene.markers());
}
for (Chromosome chr : snpEffectPredictor.getGenome()) markers.add(chr);
// ---
// Calculate raw counts
// ---
for (Marker m : markers) {
String mtype = markerTypes.getType(m);
String msubtype = markerTypes.getSubType(m);
rawCountMarkers.inc(mtype);
rawCountBases.inc(mtype, m.size());
// Count sub-types (if any)
if (msubtype != null) {
rawCountMarkers.inc(msubtype);
rawCountBases.inc(msubtype, m.size());
}
}
// ---
// Count number of bases for each marker type (overlap and join)
// ---
for (String mtype : rawCountMarkers.keysSorted()) {
if (mtype.equals(Chromosome.class.getSimpleName()))
continue; // We calculate chromosomes later (it's faster)
if (verbose) System.err.print(mtype + ":");
if (countMarkers.get(mtype) == 0) {
for (Chromosome chr : snpEffectPredictor.getGenome()) countBases(mtype, chr, markers);
}
if (verbose) System.err.println("");
}
// Show chromosomes length
String mtype = Chromosome.class.getSimpleName();
for (Chromosome chr : snpEffectPredictor.getGenome()) {
countBases.inc(mtype, chr.size());
countMarkers.inc(mtype);
}
}
public void testCase_05_PaeruPA14muccA() {
String genome = "paeru.PA14";
String gff3File = "tests/paeru.PA14.muccA.gff";
String resultFile = "tests/paeru.PA14.muccA.txt";
SnpEffectPredictor sep = buildAndCompare(genome, gff3File, resultFile, true, false);
// Make sure no splice site is added
Gene gene = sep.getGenome().getGenes().iterator().next();
Transcript tr = gene.iterator().next();
List<SpliceSite> spliceSites = tr.createSpliceSites(SpliceSite.CORE_SPLICE_SITE_SIZE, 0, 0, 0);
Assert.assertEquals(0, spliceSites.size());
}
/**
* Run main analysis
*
* @param xmlFileName
*/
@Override
public boolean run() {
// Initialzie
readConfig(); // Read config file
if (verbose)
Timer.showStdErr(
"Reading database for genome version '"
+ genomeVer
+ "' from file '"
+ config.getFileSnpEffectPredictor()
+ "' (this might take a while)");
SnpEffectPredictor snpEffectPredictor = config.loadSnpEffectPredictor();
genome = config.getGenome();
if (verbose) Timer.showStdErr("done");
// Build transcript map
for (Gene gene : snpEffectPredictor.getGenome().getGenes())
for (Transcript tr : gene) trById.put(tr.getId(), tr);
// Parse all XML files in directory
String files[] = (new File(xmlDirName)).list();
for (String xmlFileName : files) {
if (xmlFileName.endsWith(".xml.gz") || xmlFileName.endsWith(".xml")) {
String path = xmlDirName + "/" + xmlFileName;
parse(path);
}
}
// Show stats
if (verbose)
Timer.showStdErr(
"Proteing sequences:" //
+ "\n\tMatch : "
+ proteinOk.size() //
+ "\n\tDifferences : "
+ proteinDifferences.size() //
+ "\n\tAA errros : "
+ aaErrors //
);
analyzeSequenceConservation();
// Save database
save();
if (verbose) Timer.showStdErr("Done!");
return true;
}
/**
* Build a genome from a GFF3 file and compare results to 'expected' results
*
* @param genome
* @param gff3File
* @param resultFile
*/
public SnpEffectPredictor buildAndCompare(
String genome,
String gff3File,
String resultFile,
boolean readSeqs,
boolean createRandSequences) {
String expectedResult = (resultFile == null ? "" : Gpr.readFile(resultFile).trim());
// Build
Config config = new Config(genome, Config.DEFAULT_CONFIG_FILE);
SnpEffPredictorFactoryGff3 fgff3 = new SnpEffPredictorFactoryGff3(config);
fgff3.setFileName(gff3File);
fgff3.setReadSequences(readSeqs);
fgff3.setCreateRandSequences(createRandSequences);
SnpEffectPredictor sep = fgff3.create();
// Compare result
String result = show(sep.getGenome()).trim();
System.out.println("Result:\n----------\n" + result + "\n----------\n");
Assert.assertEquals(Gpr.noSpaces(expectedResult), Gpr.noSpaces(result));
return sep;
}