comparison java-genomics-toolkit/src/edu/unc/genomics/ngs/BaseAlignCounts.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.ngs;
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.io.IntervalFile;
19
20 public class BaseAlignCounts extends CommandLineTool {
21
22 private static final Logger log = Logger.getLogger(BaseAlignCounts.class);
23
24 @Parameter(names = {"-i", "--input"}, description = "Input file (reads)", required = true)
25 public IntervalFile<? extends Interval> intervalFile;
26 @Parameter(names = {"-a", "--assembly"}, description = "Genome assembly", required = true)
27 public Assembly assembly;
28 @Parameter(names = {"-x", "--extend"}, description = "Extend reads from 5' end (default = read length)")
29 public Integer extend;
30 @Parameter(names = {"-o", "--output"}, description = "Output file (Wig)", required = true)
31 public Path outputFile;
32
33 @Override
34 public void run() throws IOException {
35 log.debug("Initializing output file");
36 int mapped = 0;
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
58 Iterator<? extends Interval> it = intervalFile.query(chr, start, stop);
59 while (it.hasNext()) {
60 Interval entry = it.next();
61 int entryStop = entry.getStop();
62 if (extend != null) {
63 if (entry.isWatson()) {
64 entryStop = entry.getStart() + extend;
65 } else {
66 entryStop = entry.getStart() - extend;
67 }
68 }
69
70 for (int i = entry.getStart(); i <= entryStop; i++) {
71 count[i-start]++;
72 }
73 mapped++;
74 }
75
76 // Write the count at each base pair to the output file
77 for (int i = 0; i < count.length; i++) {
78 writer.write(count[i]);
79 writer.newLine();
80 }
81
82 // Process the next chunk
83 start = stop + 1;
84 }
85 }
86 }
87
88 log.info("Mapped "+mapped+" reads");
89 }
90
91 public static void main(String[] args) {
92 new BaseAlignCounts().instanceMain(args);
93 }
94 }