source: branches/0.13-stable/test/pileup.cc @ 3431

Last change on this file since 3431 was 3431, checked in by Peter, 6 years ago

revert r3426 and some cosmetic changes (refs #794)

  • Property svn:eol-style set to native
  • Property svn:keywords set to Id
File size: 8.8 KB
Line 
1// $Id: pileup.cc 3431 2015-10-26 07:09:34Z peter $
2
3/*
4  Copyright (C) 2014, 2015 Peter Johansson
5
6  This file is part of the yat library, http://dev.thep.lu.se/yat
7
8  The yat library is free software; you can redistribute it and/or
9  modify it under the terms of the GNU General Public License as
10  published by the Free Software Foundation; either version 3 of the
11  License, or (at your option) any later version.
12
13  The yat library is distributed in the hope that it will be useful,
14  but WITHOUT ANY WARRANTY; without even the implied warranty of
15  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16  General Public License for more details.
17
18  You should have received a copy of the GNU General Public License
19  along with yat. If not, see <http://www.gnu.org/licenses/>.
20*/
21
22#include <config.h>
23
24#include "Suite.h"
25
26#include <yat/utility/config_public.h>
27
28#ifdef YAT_HAVE_LIBBAM
29#include <yat/omic/BamFile.h>
30#include <yat/omic/BamReadIterator.h>
31#include "yat/omic/Pileup.h"
32#endif
33
34#include <yat/utility/Aligner.h>
35
36#include <algorithm>
37#include <cassert>
38#include <string>
39#include <vector>
40
41using namespace theplu::yat;
42
43// The simplest of genotypers; reporting the first and second "thirdiles"
44std::string genotype(std::string str)
45{
46  std::string res("  ");
47  std::sort(str.begin(), str.end());
48  res[0] = str[str.size()/3];
49  res[1] = str[2*str.size()/3];
50  return res;
51}
52
53
54char seq_nt16(size_t x)
55{
56#if YAT_HAVE_HTSLIB
57  return seq_nt16_str[x];
58#elif HAVE_BAM_NT16_REV_TABLE
59  return bam_nt16_rev_table[x];
60#else
61  std::string table = "=ACMGRSVTWYHKDBN";
62  return table[c];
63#endif
64}
65
66#ifdef YAT_HAVE_LIBBAM
67void test1(test::Suite& suite, const std::string& fn)
68{
69  omic::InBamFile in;
70  in.open(fn);
71
72  using omic::BamReadIterator;
73  BamReadIterator first(in);
74  BamReadIterator last;
75
76  size_t count = 0;
77  std::vector<std::string> correct_gt;
78  correct_gt.push_back("AA");
79  correct_gt.push_back("CC");
80  correct_gt.push_back("CC");
81  correct_gt.push_back("TT");
82  correct_gt.push_back("=A");
83  correct_gt.push_back("AA");
84  correct_gt.push_back("AA");
85  correct_gt.push_back("GG");
86  correct_gt.push_back("AA");
87  correct_gt.push_back("AA");
88  correct_gt.push_back("AA");
89
90  correct_gt.push_back("GG");
91  correct_gt.push_back("GG");
92  correct_gt.push_back("CC");
93  correct_gt.push_back("CC");
94  correct_gt.push_back("CC");
95
96  correct_gt.push_back("TT");
97  correct_gt.push_back("TT");
98  correct_gt.push_back("=G");
99  correct_gt.push_back("=C");
100
101  using omic::Pileup;
102  Pileup<BamReadIterator> pileup(first, last);
103  while (pileup.good()) {
104    if (pileup.pos()+1 < 24341) {
105      pileup.increment();
106      continue;
107    }
108    Pileup<BamReadIterator>::const_iterator iter = pileup.begin();
109    std::string str;
110    for (; iter!=pileup.end(); ++iter) {
111      if (iter->bam().end() <= pileup.pos()) {
112        suite.err() << "error: " << pileup.pos() << " "
113                    << iter->bam().pos() << " "
114                    << iter->bam().end() << " "
115                    << iter->bam().cigar_str()
116                    << "\n";
117        suite.add(false);
118        continue;
119      }
120      str += seq_nt16(iter->sequence());
121    }
122
123    if (pileup.pos()+1 == 24341 && count!=0) {
124      suite.err() << "incorrect count\n";
125      suite.add(false);
126    }
127
128    if (count < correct_gt.size())
129      if (genotype(str) != correct_gt[count]) {
130        suite.err() << "incorrect genotype: " << genotype(str) << "\n";
131        suite.err() << "expected: " << correct_gt[count] << "\n";
132        suite.add(false);
133      }
134
135    if (count <= 3) {
136      if (pileup.pos()+1 != static_cast<int>(count+24341)) {
137        suite.err() << "incorrect pos or count\n";
138        suite.err() << "pos: " << pileup.pos()+1 << "\n";
139        suite.err() << "count: " << count << "\n";
140        suite.add(false);
141      }
142    }
143    else if (count == 4 && !pileup.skip_ref() ) {
144      suite.out() << "expected skip ref\n";
145      suite.add(false);
146    }
147    else if (count==5) {
148      if (pileup.pos()+1 != static_cast<int>(count+24340)) {
149        suite.err() << "incorrect pos or count\n";
150        suite.err() << "pos: " << pileup.pos()+1 << "\n";
151        suite.err() << "count: " << count << "\n";
152        suite.add(false);
153      }
154    }
155    pileup.increment();
156    ++count;
157  }
158}
159
160
161void test2(test::Suite& suite, const std::string& fn)
162{
163  std::ostringstream error;
164  theplu::yat::omic::InBamFile in(fn);
165  using omic::BamRead;
166  using omic::BamReadIterator;
167  using omic::Pileup;
168  BamReadIterator first(in, 1, 24150, 25000);
169  BamReadIterator last;
170
171  std::vector<BamRead> reads(2);
172  reads[0] = *first;
173  ++first;
174  reads[1] = *first;
175  in.close();
176  std::string ref = reads[0].sequence();
177
178  suite.out() << "no modifications\n";
179  // no modification
180  typedef std::vector<BamRead>::iterator iterator;
181  typedef Pileup<iterator>::const_iterator plp_iterator;
182  for (Pileup<iterator> plp(reads.begin(), reads.end());plp.pos()<24070;
183       plp.increment()) {
184    for (plp_iterator i = plp.begin(); i!=plp.end(); ++i) {
185      if (i->bam().pos()!=reads[1].pos())
186        continue;
187      char nt = seq_nt16(i->sequence());
188      if (nt != ref[plp.pos()-reads[0].pos()]) {
189        error << "'" << nt << "' not equal '"
190              << ref[plp.pos()-reads[0].pos()]
191              << "' at pos " << reads[0].pos() << " for bam with start pos: "
192              << i->bam().pos();
193        throw std::runtime_error(error.str());
194      }
195    }
196  }
197
198  // insertion
199  suite.out() << "insertion\n";
200  assert(reads[1].sequence_length()==100);
201  utility::Aligner::Cigar cig;
202  cig.push_back(BAM_CMATCH, 10);
203  cig.push_back(BAM_CINS, 1);
204  cig.push_back(BAM_CMATCH, 89);
205  reads[1].cigar(cig);
206  int count = 0;
207  for (Pileup<iterator> plp(reads.begin(), reads.end());plp.good();
208       plp.increment()) {
209    for (plp_iterator i = plp.begin(); i!=plp.end(); ++i) {
210      if (!same_query_name(i->bam(), reads[1]))
211        continue;
212      ++count;
213      suite.err() << count << "\n";
214      char nt = seq_nt16(i->sequence());
215      if (nt != reads[1].sequence()[count-1])
216        error << "nt: " << nt << " not as expected. count: " << count << "\n";
217      // first 10 matches
218      if (count <= 10) {
219        if (plp.skip_ref())
220          error << count << " unexpected skip_ref\n";
221        if (plp.pos() != reads[1].pos()+count-1)
222          error << "count: " << count << " with pos: " << plp.pos() << "\n";
223      }
224      else if (count==11) { // insertion
225        if (!plp.skip_ref())
226          error << count << " unexpected skip_ref\n";
227      }
228      else {
229        if (plp.skip_ref())
230          error << count << " unexpected skip_ref\n";
231        if (plp.pos() != reads[1].pos()+count-2)
232          error << "count: " << count << " with pos: " << plp.pos() << "\n";
233      }
234
235      if (error.str().size())
236        throw std::runtime_error(error.str());
237    }
238  }
239
240
241  // deletion
242  suite.out() << "deletion\n";
243  assert(reads[1].sequence_length()==100);
244  cig.clear();
245  cig.push_back(BAM_CMATCH, 10);
246  cig.push_back(BAM_CDEL, 1);
247  cig.push_back(BAM_CMATCH, 90);
248  assert(cig.size()==3);
249  reads[1].cigar(cig);
250  count = 0;
251  for (Pileup<iterator> plp(reads.begin(), reads.end()); plp.good();
252       plp.increment()) {
253    for (plp_iterator i = plp.begin(); i!=plp.end(); ++i) {
254      if (!same_query_name(i->bam(), reads[1]))
255        continue;
256      ++count;
257      char nt = seq_nt16(i->sequence());
258      if (plp.pos() != reads[1].pos()+count-1)
259        error << "count: " << count << " with pos: " << plp.pos() << "\n";
260      // first 10 matches
261      if (count <= 10) {
262        if (plp.skip_ref())
263          error << count << " unexpected skip_ref\n";
264        if (nt != reads[1].sequence()[count-1])
265          error << "nt: " << nt << " not as expected\n";
266      }
267      else if (count==11) {
268        if (plp.skip_ref())
269          error << count << " unexpected skip_ref\n";
270        if (!i->is_del())
271          error << count << " expected deletion\n";
272      }
273      else {
274        if (plp.skip_ref())
275          error << count << " unexpected skip_ref\n";
276        if (nt != reads[1].sequence()[count-2])
277          error << "nt: " << nt << " not as expected\n";
278      }
279
280      if (error.str().size())
281        throw std::runtime_error(error.str());
282    }
283  }
284
285  // soft clipped
286  suite.out() << "soft clipped\n";
287  assert(reads[1].sequence_length()==100);
288  cig.clear();
289  cig.push_back(BAM_CSOFT_CLIP, 10);
290  cig.push_back(BAM_CMATCH, 90);
291  reads[1].cigar(cig);
292  count = 0;
293  for (Pileup<iterator> plp(reads.begin(), reads.end()); plp.good();
294       plp.increment()) {
295    for (plp_iterator i = plp.begin(); i!=plp.end(); ++i) {
296      if (!same_query_name(i->bam(), reads[1]))
297        continue;
298      ++count;
299      char nt = seq_nt16(i->sequence());
300      if (plp.skip_ref())
301        error << count << " unexpected skip_ref\n";
302      if (plp.pos() != reads[1].pos()+count-1)
303        error << "count: " << count << " with pos: " << plp.pos() << "\n";
304      if (nt != reads[1].sequence()[count+9])
305        error << "count: " << count << " nt: " << nt << " not as expected\n";
306
307      if (error.str().size())
308        throw std::runtime_error(error.str());
309    }
310  }
311}
312#else
313// tests if libbam is not available, should never be run
314void test1(test::Suite& suite, const std::string& fn) {assert(0);}
315void test2(test::Suite& suite, const std::string& fn) {assert(0);}
316#endif
317
318int main(int argc, char* argv[])
319{
320  test::Suite suite(argc, argv, true);
321  std::string fn = "../../data/foo.sorted.bam";
322  try {
323    test1(suite, fn);
324    test2(suite, fn);
325  }
326  catch (std::runtime_error& e) {
327    suite.err() << "test failed: " << e.what() << "\n";
328    suite.add(false);
329  }
330  return suite.return_value();
331}
Note: See TracBrowser for help on using the repository browser.