Mercurial > repos > timpalpant > java_genomics_toolkit
comparison src/edu/unc/genomics/nucleosomes/MapDyads.java @ 2:e16016635b2a
Uploaded
author | timpalpant |
---|---|
date | Mon, 13 Feb 2012 22:12:06 -0500 |
parents | |
children |
comparison
equal
deleted
inserted
replaced
1:a54db233ee3d | 2:e16016635b2a |
---|---|
1 package edu.unc.genomics.nucleosomes; | |
2 | |
3 import java.io.BufferedWriter; | |
4 import java.io.IOException; | |
5 import java.nio.charset.Charset; | |
6 import java.nio.file.Files; | |
7 import java.nio.file.Path; | |
8 import java.util.Iterator; | |
9 | |
10 import org.apache.log4j.Logger; | |
11 | |
12 import com.beust.jcommander.Parameter; | |
13 | |
14 import edu.ucsc.genome.TrackHeader; | |
15 import edu.unc.genomics.Assembly; | |
16 import edu.unc.genomics.CommandLineTool; | |
17 import edu.unc.genomics.Interval; | |
18 import edu.unc.genomics.PositiveIntegerValidator; | |
19 import edu.unc.genomics.io.IntervalFile; | |
20 | |
21 public class MapDyads extends CommandLineTool { | |
22 | |
23 private static final Logger log = Logger.getLogger(MapDyads.class); | |
24 | |
25 @Parameter(names = {"-i", "--input"}, description = "Input file (reads)", required = true) | |
26 public IntervalFile<? extends Interval> inputFile; | |
27 @Parameter(names = {"-s", "--size"}, description = "Mononucleosome length (default: read length)", validateWith = PositiveIntegerValidator.class) | |
28 public Integer nucleosomeSize; | |
29 @Parameter(names = {"-a", "--assembly"}, description = "Genome assembly", required = true) | |
30 public Assembly assembly; | |
31 @Parameter(names = {"-o", "--output"}, description = "Output file (Wig)", required = true) | |
32 public Path outputFile; | |
33 | |
34 @Override | |
35 public void run() throws IOException { | |
36 log.debug("Initializing output file"); | |
37 int mapped = 0; | |
38 try (BufferedWriter writer = Files.newBufferedWriter(outputFile, Charset.defaultCharset())) { | |
39 // Write the Wiggle track header to the output file | |
40 TrackHeader header = new TrackHeader("wiggle_0"); | |
41 header.setName("Converted " + inputFile.getPath().getFileName()); | |
42 header.setDescription("Converted " + inputFile.getPath().getFileName()); | |
43 writer.write(header.toString()); | |
44 writer.newLine(); | |
45 | |
46 // Process each chromosome in the assembly | |
47 for (String chr : assembly) { | |
48 log.debug("Processing chromosome " + chr); | |
49 // Write the contig header to the output file | |
50 writer.write("fixedStep chrom="+chr+" start=1 step=1 span=1"); | |
51 writer.newLine(); | |
52 | |
53 int start = 1; | |
54 while (start < assembly.getChrLength(chr)) { | |
55 int stop = start + DEFAULT_CHUNK_SIZE - 1; | |
56 int length = stop - start + 1; | |
57 int[] count = new int[length]; | |
58 | |
59 Iterator<? extends Interval> it = inputFile.query(chr, start, stop); | |
60 while (it.hasNext()) { | |
61 Interval entry = it.next(); | |
62 if (nucleosomeSize == null) { | |
63 count[entry.center()-start]++; | |
64 } else { | |
65 count[entry.getStart()+nucleosomeSize-start]++; | |
66 } | |
67 mapped++; | |
68 } | |
69 | |
70 // Write the average at each base pair to the output file | |
71 for (int i = 0; i < count.length; i++) { | |
72 writer.write(count[i]); | |
73 writer.newLine(); | |
74 } | |
75 | |
76 // Process the next chunk | |
77 start = stop + 1; | |
78 } | |
79 } | |
80 } | |
81 | |
82 log.info("Mapped "+mapped+" reads"); | |
83 } | |
84 | |
85 public static void main(String[] args) { | |
86 new MapDyads().instanceMain(args); | |
87 } | |
88 } |