public String getOutputRecord(int tagIndex, double[][] theResults) {
int GNum = 0;
for (int i = 0; i < theResults.length; i++) {
if (theResults[i] != null) GNum++;
}
StringBuilder sb = new StringBuilder();
sb.append(BaseEncoder.getSequenceFromLong(tbt.getTag(tagIndex))).append("\t");
sb.append(String.valueOf(tbt.getNumberOfTaxaWithTag(tagIndex))).append("\t");
sb.append(String.valueOf(GNum)).append("\t");
// chr, bestSite, refAlignment.getPositionInLocus(bestSite), bestP, countSig, (double)-1,
// bestNewTestFamilies.length, (double)this.getFamilyCode(bestNewTestFamilies)}
for (int i = 0; i < theResults.length; i++) {
if (theResults[i] == null) continue;
sb.append(String.valueOf((int) theResults[i][0])).append("\t");
sb.append(String.valueOf((int) theResults[i][1])).append("\t");
sb.append(String.valueOf((int) theResults[i][2])).append("\t");
sb.append(String.valueOf(theResults[i][3])).append("\t");
sb.append(String.valueOf((int) theResults[i][4])).append("\t");
sb.append(String.valueOf(theResults[i][5])).append("\t");
sb.append(String.valueOf((int) theResults[i][6])).append("\t");
sb.append(String.valueOf((long) theResults[i][7])).append("\t");
}
sb.append("\n");
return sb.toString();
}
private long[] getTagsInBits(TagsByTaxaByte aTBT, int tagIndex, int[] reDirect, int anchorTaxa) {
int lgPerSite = (anchorTaxa / 64) + 1;
long[] seq = new long[lgPerSite];
for (int j = 0; j < aTBT.getTaxaCount(); j++) {
if (reDirect[j] < 0) continue;
int index = reDirect[j] / 64;
int offset = reDirect[j] % 64;
if (aTBT.getReadCountForTagTaxon(tagIndex, j) > 0) { // reference alleles
seq[index] = seq[index] | (1L << offset);
}
}
return seq;
}
public void run() {
long starttime = System.currentTimeMillis();
ArrayList<String> resultList = new ArrayList();
for (int i = 0; i < tbt.getTagCount(); i++) {
int[] minPresentFamilies = getMinPresentFamily(i, (float) 0.08);
if (minPresentFamilies == null) continue;
ArrayList segregationList =
this.getSegregationFamilyGroups(i, minPresentFamilies, 10, pThresh);
if (segregationList == null) continue;
int[][] segregationFamilyGroups = (int[][]) segregationList.get(0);
int[] bestChrs = (int[]) segregationList.get(1);
int minGnum = segregationFamilyGroups.length;
if (familyGroupNum < minGnum) minGnum = familyGroupNum;
double[][] theResults = new double[minGnum][];
boolean nullFlag = true;
for (int j = 0;
j < minGnum;
j++) { // it should be bestChrs.length, only caculate 2 loci for now
ScanChromosome scanOnBestChrom =
new ScanChromosome(
tbt,
i,
theAnchorChrInBits[bestChrs[j] - 1],
segregationFamilyGroups[j],
j,
theResults,
0.00001,
1);
scanOnBestChrom.scan();
if (theResults[j] == null) continue;
nullFlag = false;
}
if (nullFlag) continue;
resultList.add(this.getOutputRecord(i, theResults));
}
result = resultList.toArray(new String[resultList.size()]);
double secs = (double) (System.currentTimeMillis() - starttime) * 0.001;
System.out.println(
"Task "
+ taskID
+ " with "
+ tbt.getTagCount()
+ " tags is finished in "
+ secs
+ " seconds");
}
public long[] getTagDistInBits(int[] newTestFamilies) {
int lgPerSite = (refAlignment.getSequenceCount() / 64) + 1;
long[] seq = new long[lgPerSite];
for (int i = 0; i < newTestFamilies.length; i++) {
for (int j = 0; j < family2TBTIndex[newTestFamilies[i]].length; j++) {
int index =
tbt2anchorRedirect[newTestFamilies[i]][family2TBTIndex[newTestFamilies[i]][j]] / 64;
int offset =
tbt2anchorRedirect[newTestFamilies[i]][family2TBTIndex[newTestFamilies[i]][j]] % 64;
if (tbt.getReadCountForTagTaxon(tagIndex, family2TBTIndex[newTestFamilies[i]][j])
> 0) { // reference alleles
seq[index] = seq[index] | (1L << offset);
}
}
}
return seq;
}
private int[] getMinPresentFamily(int tagIndex, float minPresentInFamily) {
ArrayList<Integer> overMinPresentFamilyList = new ArrayList();
for (int i = 0; i < family2TBTIndex.length; i++) {
int cnt = 0;
for (int j = 0; j < family2TBTIndex[i].length; j++) {
if (tbt.getReadCountForTagTaxon(tagIndex, family2TBTIndex[i][j]) != 0) cnt++;
}
float rate = (float) cnt / family2TBTIndex[i].length;
if (rate > minPresentInFamily) overMinPresentFamilyList.add(i);
}
if (overMinPresentFamilyList.isEmpty()) return null;
int[] overMinPresentFamily = new int[overMinPresentFamilyList.size()];
for (int i = 0; i < overMinPresentFamily.length; i++) {
overMinPresentFamily[i] = overMinPresentFamilyList.get(i);
}
return overMinPresentFamily;
}
public void run() {
long starttime = System.currentTimeMillis();
ArrayList<String> resultList = new ArrayList();
for (int i = 0; i < tbt.getTagCount(); i++) {
if (tbt.getNumberOfTaxaWithTag(i) < 30) continue;
double[] bestR = {-1, -1, -1, 1, -1};
int blastChr = -1, blastPos = -1, bestAlignment = -1, refDiv = -1;
long[] testTag = tbt.getTag(i);
long[] testTagDist;
double[][] theResults = new double[theAnchorChrInBits.length][];
ScanChromosomeMTR[] scanOnChr = new ScanChromosomeMTR[theAnchorChrInBits.length];
for (int cn = 0; cn < theAnchorChrInBits.length; cn++) {
testTagDist =
getTagsInBits(
tbt, i, tbt2anchRedirect[cn], theAnchorChrInBits[cn].getSequenceCount());
scanOnChr[cn] =
new ScanChromosomeMTR(
theAnchorChrInBits[cn], testTagDist, cn, theResults, 0.000001, 1, blastPos);
scanOnChr[cn].scan();
}
int countRealSig = 0;
double[] pRank = new double[theAnchorChrInBits.length];
for (int cn = 0; cn < theAnchorChrInBits.length; cn++) {
double[] r = theResults[cn];
pRank[cn] = r[3];
if (r[3] < bestR[3]) {
bestR = r.clone();
bestAlignment = cn;
}
if (r[3] < 0.000001) countRealSig++;
}
Arrays.sort(pRank);
double[][] bestResWithNewThreshold = new double[1][];
testTagDist =
getTagsInBits(
tbt,
i,
tbt2anchRedirect[bestAlignment],
theAnchorChrInBits[bestAlignment].getSequenceCount());
ScanChromosomeMTR bestChrNewThres =
new ScanChromosomeMTR(
theAnchorChrInBits[bestAlignment],
testTagDist,
0,
bestResWithNewThreshold,
pRank[1],
1,
blastPos);
bestChrNewThres.scan();
int countOfSitesBetterThanNextBestChr = (int) bestResWithNewThreshold[0][4];
String s =
String.format(
"%s %d %d %d %d %d %d %d %g %d %d %d %g %g %d %d %d %n",
BaseEncoder.getSequenceFromLong(testTag),
i,
blastChr,
blastPos,
refDiv,
(int) bestR[0],
(int) bestR[1],
(int) bestR[2],
bestR[3],
(int) bestR[4],
tbt.getNumberOfTaxaWithTag(i),
countRealSig,
Math.log10(pRank[1] / pRank[0]),
Math.log10(pRank[theAnchorChrInBits.length / 2] / pRank[0]),
countOfSitesBetterThanNextBestChr,
bestChrNewThres.minSigPos,
bestChrNewThres.maxSigPos);
resultList.add(s);
}
result = resultList.toArray(new String[resultList.size()]);
double secs = (double) (System.currentTimeMillis() - starttime) * 0.001;
System.out.println(
"Task "
+ taskID
+ " with "
+ tbt.getTagCount()
+ " tags is finished in "
+ secs
+ " seconds");
}