/**
* find the kmers in the contigs
*
* @param reader
* @throws IOException
*/
private void processContigFile(SequenceReader reader) throws IOException {
Sequence seq;
NuclKmerGenerator kmerGenerator;
Kmer kmer;
int contigIdx = 0;
while ((seq = reader.readNextSequence()) != null) {
if (seq.getSeqString().length() < kmerSize) {
continue;
}
// use int to represent seqname in case contig names are too long
contigMap.put(
contigIdx, new Contig(seq.getSeqName(), seq.getSeqString().length() - kmerSize + 1));
// forward direction
kmerGenerator = new NuclKmerGenerator(seq.getSeqString(), kmerSize);
while (kmerGenerator.hasNext()) {
kmer = kmerGenerator.next();
KmerAbund kmerAbund = kmerMaps[0].get(kmer);
if (kmerAbund == null) {
kmerAbund = new KmerAbund();
kmerMaps[0].put(kmer, kmerAbund);
}
kmerAbund.contigList.add(new ContigCoverage(contigIdx, kmerGenerator.getPosition() - 1));
}
// reverse direction
kmerGenerator =
new NuclKmerGenerator(IUBUtilities.reverseComplement(seq.getSeqString()), kmerSize);
while (kmerGenerator.hasNext()) {
kmer = kmerGenerator.next();
KmerAbund kmerAbund = kmerMaps[1].get(kmer);
if (kmerAbund == null) {
kmerAbund = new KmerAbund();
kmerMaps[1].put(kmer, kmerAbund);
}
kmerAbund.contigList.add(
new ContigCoverage(
contigIdx,
seq.getSeqString().length() - kmerGenerator.getPosition() - kmerSize + 1));
}
contigIdx++;
}
reader.close();
}
public void printCovereage(OutputStream coverage_out, OutputStream abundance_out)
throws IOException {
adjustCount();
// print out the weighted kmer coverage
// we found mean coverage matched the previous biological observation
PrintStream coverage_outStream = new PrintStream(coverage_out);
coverage_outStream.println("#total reads: " + totalReads.intValue());
coverage_outStream.println("#use mean_cov to adjust the contig abundance, not median_cov ");
coverage_outStream.println(
"#seqid\tmean_cov\tmedian_cov\ttotal_pos\tcovered_pos\tcovered_ratio");
for (Contig contig : contigMap.values()) {
ArrayList<Double> counts = new ArrayList<Double>();
int coveredPos = 0;
for (int pos = 0; pos < contig.coverage.length; pos++) {
if (contig.coverage[pos] > 0) {
coveredPos++;
}
counts.add(contig.coverage[pos]);
}
if (coveredPos > 0) {
coverage_outStream.println(
contig.name
+ "\t"
+ String.format(dformat, StdevCal.calMean(counts))
+ "\t"
+ String.format(dformat, (StdevCal.calMedian(counts)))
+ "\t"
+ counts.size()
+ "\t"
+ coveredPos
+ "\t"
+ String.format(dformat, (double) coveredPos / (double) contig.coverage.length));
} else { // no coverage
coverage_outStream.println(
contig.name
+ "\t"
+ 0
+ "\t"
+ 0
+ "\t"
+ contig.coverage.length
+ "\t"
+ 0
+ "\t"
+ 0);
}
}
coverage_outStream.close();
// print kmer abundance
HashMap<Integer, Integer> abundanceCountMap =
new HashMap<Integer, Integer>(); // the frequeny of the kmer abundance
PrintStream abundance_outStream = new PrintStream(abundance_out);
// need to merge the counts from forward and reverse together.
HashSet<Kmer> kmerSet = new HashSet<Kmer>();
kmerSet.addAll(kmerMaps[0].keySet());
for (Kmer kmer : kmerSet) {
AtomicInteger abundance = kmerMaps[0].get(kmer).count;
String reverseKmerStr = IUBUtilities.reverseComplement(kmer.decodeLong(kmer.getLongKmers()));
Kmer reverseKmer = (new NuclKmerGenerator(reverseKmerStr, this.kmerSize)).next();
KmerAbund kmerAbund = kmerMaps[1].get(reverseKmer);
if (kmerAbund != null) {
abundance.addAndGet(kmerAbund.count.get());
}
Integer count = abundanceCountMap.get(abundance.get());
if (count == null) {
abundanceCountMap.put(abundance.get(), 1);
} else {
abundanceCountMap.put(abundance.get(), count + 1);
}
}
abundance_outStream.println("kmer_abundance\tfrequency");
for (Integer abundance : abundanceCountMap.keySet()) {
abundance_outStream.println(abundance + "\t" + abundanceCountMap.get(abundance));
}
abundance_outStream.close();
}