protected MnaseOutput countMnaseReads(
String chr,
int pos,
ChromScoresFast fwCounts,
ChromScoresFast revCounts,
boolean chipSeqMetric)
throws Exception {
double PSEUDOCOUNT = 0.005; // per bp
int FRAGLENHIGH = 600;
int center = pos;
int quarterNuc = (int) ((double) this.footprintSize / 4.0);
int halfNuc = (int) ((double) this.footprintSize / 2.0);
int assocHalf = (int) ((double) this.assocSize / 2.0);
int fraghalf = (int) ((double) FRAGLENHIGH / 2.0);
int cycle = halfNuc;
int s, e;
MnaseOutput out = new MnaseOutput();
// int windHalf = (int)((double)this.normalizationWindow/2.0);
// s = center - windHalf;
// e = center + windHalf;
// out.normWindCount = (double)this.countUnstrandedReads(fwCounts, revCounts,chr, s, e);
// // System.err.printf("NormWind Cpg %d\t Getting (all) reads %s at %d-%d\t%d\n", center,
// mnasePrefix, s, e,(int)normWindCount);
if (chipSeqMetric) {
// CHIP-SEQ . THIS SHOULD TAKE A FRAGMENT LEN, BUT 600 is based on K4 IMR90 offsets.
System.err.println("Why are we using ChIP-seq metric??");
System.exit(1);
// First fw
s = center - fraghalf;
e = center;
int preCountPos = countReads(fwCounts, chr, s, e);
// if (preCountPos>0) System.err.printf("Cpg %d\t Getting (+) reads at %d-%d\t%d\n", center,
// s, e,preCountPos);
// Then rev
s = center;
e = center + fraghalf;
int postCountNeg = countReads(revCounts, chr, s, e);
// if (postCountNeg>0) System.err.printf("Cpg %d\t Getting (-) reads at %d-%d\t%d\n", center,
// s, e,postCountNeg);
out.val = (double) (preCountPos + postCountNeg) / (double) fraghalf;
out.rawCounts = preCountPos + postCountNeg;
// if (this.normalizationWindow>0.0)
// {
// out = ((double)(preCountPos+postCountNeg)+(PSEUDOCOUNT*(double)fraghalf)) *
// (1.0/((double)normWindCount+(PSEUDOCOUNT*(double)this.normalizationWindow))) *
// (((double)this.normalizationWindow)/((double)fraghalf));
// out = Math.log(out)/LOG_OF_TWO;
// }
} else {
// MNASE METRIC
// First fw
s = center - cycle - assocHalf;
e = center - cycle + assocHalf;
int preCountPos = countReads(fwCounts, chr, s, e);
// System.err.printf("Cpg %d\t Getting (+) reads at %d-%d\t%d\n", center, s, e,preCountPos);
// Then rev
s = center + cycle - assocHalf;
e = center + cycle + assocHalf;
int postCountNeg = countReads(revCounts, chr, s, e);
// System.err.printf("Cpg %d\t Getting (-) reads at %d-%d\t%d\n", center, s, e,postCountNeg);
int phasedCount = preCountPos + postCountNeg;
int phasedLen = 2 * assocHalf;
out.rawCounts = phasedCount;
out.val = phasedCount;
// Then unphased
int unphasedCount = 0;
s = center - assocHalf;
e = center + assocHalf;
unphasedCount += countUnstrandedReads(fwCounts, revCounts, chr, s, e);
int unphasedLen = 2 * assocHalf;
// s = center - this.PERIODICITY - assocHalf;
// e = center - this.PERIODICITY + assocHalf;
// unphasedCount += countUnstrandedReads(fwCounts, revCounts, chr, s, e);
// s = center + this.PERIODICITY - assocHalf;
// e = center + this.PERIODICITY + assocHalf;
// unphasedCount += countUnstrandedReads(fwCounts, revCounts, chr, s, e);
// int unphasedLen = 6 * assocHalf;
// // System.err.printf("Cpg %d\t Getting unphased reads at %d-%d\t%d\n", center, center -
// this.PERIODICITY - assocHalf,
// // center + this.PERIODICITY + assocHalf,unphasedCount);
//
//
//
out.val =
((double) (phasedCount) + (PSEUDOCOUNT * (double) phasedLen))
* (1.0 / ((double) unphasedCount + (PSEUDOCOUNT * (double) unphasedLen)))
* (((double) unphasedLen) / ((double) phasedLen));
out.val = Math.log(out.val) / LOG_OF_TWO;
out.val = phasedCount - unphasedCount;
out.val = Math.min(preCountPos, postCountNeg) - unphasedCount;
}
return out;
}
public void doMain(String[] args) throws Exception {
CmdLineParser parser = new CmdLineParser(this);
// if you have a wider console, you could increase the value;
// here 80 is also the default
parser.setUsageWidth(80);
try {
parser.parseArgument(args);
if ((arguments.size() != 0)) {
System.err.println(C_USAGE);
parser.printUsage(System.err);
System.exit(1);
}
if ((outPrefix == null)) {
System.err.println("Must specify outPrefix and at least one mnasePrefix");
parser.printUsage(System.err);
System.err.println(C_USAGE);
System.exit(1);
}
} catch (CmdLineException e) {
System.err.println(e.getMessage());
System.err.println(C_USAGE);
// print the list of available options
parser.printUsage(System.err);
System.err.println();
return;
}
Logger.getLogger(Logger.GLOBAL_LOGGER_NAME).setLevel(Level.SEVERE);
// Setup output files and print domain finders
ListUtils.setDelim("-");
String mnasePrefixStr = mnasePrefix;
String normString =
(this.normalizationWindow == 0)
? ""
: String.format(".normalized%dbpWind", this.normalizationWindow);
String motifString =
(this.motif == null) ? "" : String.format(".motif%s", this.motif.toUpperCase());
String withinStr =
(this.withinFeat == null) ? "" : String.format(".withinFeat-%s", this.withinFeat);
if (this.filterByMotif != null)
motifString +=
String.format(
".filteredByMotif%s%s%s",
this.filterByMotif,
(this.filterByMotifInvert) ? "-inverted" : "",
(this.filterByMotifOffset != 0)
? String.format("-offset%d", this.filterByMotifOffset)
: "");
String fwOnlyStr = (this.fwStrandOnly) ? ".fwStrandOnly" : "";
String name =
String.format(
"nucleosomeReads.%s%s%s.minReads%.2f%s.nuc%d.assoc%d%s.%s.%s",
this.outPrefix,
motifString,
withinStr,
this.minReadsPerBp,
fwOnlyStr,
this.footprintSize,
this.assocSize,
normString,
mnasePrefixStr,
methPrefix);
String outFn =
String.format(
"%s%s%s.minReads%.2f%s.nuc%d.assoc%d%s.%s.%s.csv",
this.outPrefix,
motifString,
withinStr,
this.minReadsPerBp,
fwOnlyStr,
this.footprintSize,
this.assocSize,
normString,
mnasePrefixStr,
methPrefix);
String outFnWig = outFn.replace(".csv", ".wig");
// PrintWriter pw = new PrintWriter(new FileOutputStream(outFn));
PrintWriter pwWig = new PrintWriter(new FileOutputStream(outFnWig));
// pw.printf("%s,%s\n","Meth", "ReadsPerBp");
pwWig.printf("track type=wiggle_0 name=\"%s\" description=\"%s\"\n", name, name);
MethylDbQuerier params = new MethylDbQuerier();
params.setMinCTreads(this.minCTreads);
params.setUseNonconversionFilter(!this.noNonconvFilter);
params.setMaxOppstrandAfrac(this.maxOppStrandAfrac);
params.setMaxNextNonGfrac(this.maxNextNonGfrac);
if (this.withinFeat != null) params.addFeatFilter(this.withinFeat, this.featFlank);
// We use the control table simply to limit the valid Cpg. This is a result of
// our incorrect loading of Lister 2009 tables, which contains many Cpgs which
// incorrectly yield a meth level of 0 at CpGs not covered in their sequencing data.
// This was an artifact of the way they published their data, where they published
// a list of methy-C positions without positions containing 0 mC reads, so we had
// to add fake positions for all Cs in the genome, and we didn't know which ones
// actually had coverage in their data.
List<String> methTables = Arrays.asList(methPrefix, controlMethPrefix);
// List<String> mnaseTables = Arrays.asList(mnasePrefix);
int nSeries = 1;
boolean useChipseq[] = new boolean[nSeries];
for (int i = 0; i < nSeries; i++) {
// Wow talk about a special case.
useChipseq[i] = !mnasePrefix.contains("mnase") && !mnasePrefix.contains("Schones");
}
double[] methCounts = new double[METHCOUNTER_LEN];
double[] methTotals = new double[METHCOUNTER_LEN];
int chromNum = 1;
for (String chr :
Arrays.asList(
"chr22")) // MethylDbUtils.CHROMS_MINUS_TWELVE) //MethylDbUtils.SMALL_CHROMS) //
// Arrays.asList("chr22")) //,"chr18","chr19","chr20")) //
{
System.err.printf("On chrom %d (%s)\n", chromNum++, chr);
String s = String.format("variableStep\tchrom=%s\n", chr);
pwWig.append(s);
int chrInt = (new ChromFeatures()).chrom_from_public_str(chr);
// Iterator uses DB connection and can use a ton of memory because
// it loads all rows at once. This stuff should really be added to iterator
// class, but until it is , just iterate here over the chromosome
int onCpg = 0;
// Get the full array
try {
{
ChromScoresFast counts[] = new ChromScoresFast[2];
ChromScoresFast meths[] = new ChromScoresFast[2];
int offs = 0; // 20000000; //0;
int offe = 30000000; // 0;
System.err.printf("offs=%d, offe=%d\n", offs, offe);
// The mnase counts are the same for both cases
counts =
MethylDbUtils.chromScoresReadCounts(
params, chr, this.mnasePrefix, this.refGenome, offs, offe);
// We can filter counts by a particular motif
if (this.filterByMotif != null) {
ChromScoresFast filterByMotifCounts[] = new ChromScoresFast[2];
filterByMotifCounts =
MethylDbUtils.chromScoresMotifCounts(
chr, this.refGenome, this.filterByMotif, offs, offe, false);
System.err.printf(
"%s, filtering + strand for motif %s. Pre-filter count: %.0f\n",
chr, this.filterByMotif, counts[0].getScoresTotal(chr));
counts[0].mask(filterByMotifCounts[0], this.filterByMotifInvert);
// counts[0].mask(filterByMotifCounts[1],this.filterByMotifInvert);
System.err.printf(
"%s, filtering + strand for motif %s. Post-filter count: %.0f\n",
chr, this.filterByMotif, counts[0].getScoresTotal(chr));
System.err.printf(
"%s, filtering - strand for motif %s. Pre-filter count: %.0f\n",
chr, this.filterByMotif, counts[1].getScoresTotal(chr));
counts[1].mask(filterByMotifCounts[1], this.filterByMotifInvert);
// counts[1].mask(filterByMotifCounts[0],this.filterByMotifInvert);
System.err.printf(
"%s, filtering - strand for motif %s. Post-filter count: %.0f\n",
chr, this.filterByMotif, counts[1].getScoresTotal(chr));
}
// The meth differ if it's a motif
if (this.autoMnase || this.autoMnaseFw || this.autoMnaseRev) {
meths = counts;
} else if (this.motif == null) {
meths =
MethylDbUtils.chromScoresMethLevels(
params, chr, this.methPrefix, this.refGenome, offs, offe);
} else {
char[] seqArr = null;
{
Sequence seq = GoldAssembly.chromSeq(this.refGenome, chr);
String seqStr = seq.seqString();
if (offe > 0) {
seqStr = seqStr.substring(offs, offe);
}
seqArr = seqStr.toUpperCase().toCharArray();
}
System.err.println("Seq length=" + seqArr.length);
ChromScoresMotifPositions all = new ChromScoresMotifPositions(this.refGenome);
System.err.printf(
"About to populate both strands for %s for motif %s\n", chr, this.motif);
StrandedFeature.Strand motifStrand =
(this.fwStrandOnly) ? StrandedFeature.POSITIVE : StrandedFeature.UNKNOWN;
all.populate(chr, this.motif, seqArr, offs, motifStrand);
meths[0] = all;
meths[1] = all;
// System.err.printf("Fw score at %d = %d\n",24009253,counts[0].getScore(chr,
// 24009253));
// WigOptions wo = new WigOptions();
// wo.f_step = 1;
// //counts[0] = counts[0].smooth(50, 30);
// counts[0].wigOutput("testFw.wig", wo);
// counts[1].wigOutput("testRev.wig", wo);
}
int minPos = counts[0].chromMinPos(chr);
System.err.println("Getting min pos: " + minPos);
int maxPos = counts[0].chromMaxPos(chr);
System.err.println("Getting max pos: " + maxPos);
for (int pos = minPos; pos < maxPos; pos += 1) // this.step)
{
boolean fwRead = (counts[0].getScore(chr, pos).intValue() >= 1);
boolean revRead = (counts[1].getScore(chr, pos).intValue() >= 1);
if (this.fwStrandOnly) revRead = false; // fwRead=false;
boolean enoughReads = true;
MnaseOutput mnaseReads = new MnaseOutput();
mnaseReads.rawCounts = (fwRead) ? 1 : 0;
if (this.minReadsPerBp > 0.0) {
if (fwRead) {
int nucCenter = pos + ((this.PERIODICITY - this.assocSize) / 2);
mnaseReads =
this.countMnaseReads(chr, nucCenter, counts[0], counts[1], useChipseq[0]);
} else if (revRead) {
int nucCenter = pos - ((this.PERIODICITY - this.assocSize) / 2);
mnaseReads =
this.countMnaseReads(chr, nucCenter, counts[0], counts[1], useChipseq[0]);
}
enoughReads = (mnaseReads.val >= this.minReadsPerBp);
}
// if (mnaseReads.rawCounts>2) System.err.printf("Raw counts=%d\n",
// (int)mnaseReads.rawCounts);
// enoughReads = true;
if (enoughReads && fwRead)
incrementMethCounts(chr, pos, false, methCounts, methTotals, meths, maxPos);
if (enoughReads && revRead)
incrementMethCounts(chr, pos, true, methCounts, methTotals, meths, maxPos);
//
if ((pos % 1E6) == 0)
System.err.printf("On pos #%d, meth=%d\n", pos, (int) mnaseReads.rawCounts);
if (enoughReads) {
// pwWig.printf("%d\t%.2f\n",pos, mnaseReads.val);
//// pwWig.printf("%d\t%.2f\t%.2f\t%s\n",pos,mnaseReads.rawCounts,
// mnaseReads.val, fwRead?"+":"-");
//
//// pw.printf("%d,%d", chrInt, pos);
////// pw.printf(",%.3f", mnaseReads.val);
//// pw.printf(",%d", (int)mnaseReads.rawCounts);
////// pw.printf(",%.3f", mnaseReads.normWindCount);
//// pw.println();
}
}
// Try to get rid of objects
counts[0] = null;
counts[1] = null;
meths[0] = null;
meths[1] = null;
System.gc();
}
} catch (Exception e) {
System.err.printf("%s\nCouldn't do region %s\n", e.toString(), chr);
e.printStackTrace();
}
}
// pw.close();
pwWig.close();
outFn = outFn.replace(".csv", ".methAlign.csv");
PrintWriter pw = new PrintWriter(new FileOutputStream(outFn));
ListUtils.setDelim(",");
pw.println(ListUtils.excelLine(methCounts));
pw.println(ListUtils.excelLine(methTotals));
pw.close();
} // Main
