comparison java-genomics-toolkit/src/edu/unc/genomics/converters/IntervalToWig.java @ 0:1daf3026d231

Upload alpha version
author timpalpant
date Mon, 13 Feb 2012 21:55:55 -0500
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:1daf3026d231
1 package edu.unc.genomics.converters;
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.Interval;
17 import edu.unc.genomics.CommandLineTool;
18 import edu.unc.genomics.ValuedInterval;
19 import edu.unc.genomics.io.IntervalFile;
20
21 public class IntervalToWig extends CommandLineTool {
22
23 private static final Logger log = Logger.getLogger(IntervalToWig.class);
24
25 @Parameter(names = {"-i", "--input"}, description = "Input file (Bed/BedGraph)", required = true)
26 public IntervalFile<? extends Interval> intervalFile;
27 @Parameter(names = {"-a", "--assembly"}, description = "Genome assembly", required = true)
28 public Assembly assembly;
29 @Parameter(names = {"-o", "--output"}, description = "Output file (Wig)", required = true)
30 public Path outputFile;
31
32 @Override
33 public void run() throws IOException {
34 log.info(intervalFile.count() + " entries in input");
35
36 log.debug("Initializing output file");
37 try (BufferedWriter writer = Files.newBufferedWriter(outputFile, Charset.defaultCharset())) {
38 // Write the Wiggle track header to the output file
39 TrackHeader header = new TrackHeader("wiggle_0");
40 header.setName("Converted " + intervalFile.getPath().getFileName());
41 header.setDescription("Converted " + intervalFile.getPath().getFileName());
42 writer.write(header.toString());
43 writer.newLine();
44
45 // Process each chromosome in the assembly
46 for (String chr : assembly) {
47 log.debug("Processing chromosome " + chr);
48 // Write the contig header to the output file
49 writer.write("fixedStep chrom="+chr+" start=1 step=1 span=1");
50 writer.newLine();
51
52 int start = 1;
53 while (start < assembly.getChrLength(chr)) {
54 int stop = start + DEFAULT_CHUNK_SIZE - 1;
55 int length = stop - start + 1;
56 int[] count = new int[length];
57 float[] sum = new float[length];
58
59 Iterator<? extends Interval> it = intervalFile.query(chr, start, stop);
60 while (it.hasNext()) {
61 ValuedInterval entry = (ValuedInterval) it.next();
62 if (entry.getValue() != null) {
63 for (int i = entry.getStart(); i <= entry.getStop(); i++) {
64 sum[i-start] += entry.getValue().floatValue();
65 count[i-start]++;
66 }
67 }
68 }
69
70 // Write the average at each base pair to the output file
71 for (int i = 0; i < sum.length; i++) {
72 if (count[i] == 0) {
73 writer.write(String.valueOf(Float.NaN));
74 } else {
75 writer.write(String.valueOf(sum[i]/count[i]));
76 }
77 writer.newLine();
78 }
79
80 // Process the next chunk
81 start = stop + 1;
82 }
83 }
84 }
85 }
86
87 public static void main(String[] args) {
88 new IntervalToWig().instanceMain(args);
89 }
90
91 }