///////////////////////////////////////////////////////////
// rev2bSnpJoin test
// case that first came up in lrc_kir where sanity check
// fails on a seemingly straightforward 2base reverse snp
//
///////////////////////////////////////////////////////////
void rev2bSnpJoinTest(CuTest *testCase)
{
CuAssertTrue(testCase, true);
SideGraph sg;
SGPosition p1a(0, 5);
SGPosition p2a(1, 0);
sg.addSequence(new SGSequence(0, 20, "seq1"));
sg.addSequence(new SGSequence(1, 2, "seq2"));
sg.addJoin(new SGJoin(SGSide(p1a, true), SGSide(p2a, false)));
vector<string> bases(2);
bases[0] = "ACCTGACCATAGGCATGGGC";
bases[1] = "TA";
vector<SGNamedPath> paths;
Side2Seq converter;
converter.init(&sg, &bases, &paths);
try
{
converter.convert();
}
catch(exception& e)
{
cerr << "Exception caught " << e.what() << endl;
CuAssertTrue(testCase, false);
}
const SideGraph* outGraph = converter.getOutGraph();
CuAssertTrue(testCase, outGraph->getNumSequences() == 4);
CuAssertTrue(testCase, outGraph->getSequence(0)->getLength() == 5);
CuAssertTrue(testCase, outGraph->getSequence(1)->getLength() == 15);
CuAssertTrue(testCase, outGraph->getSequence(2)->getLength() == 1);
CuAssertTrue(testCase, outGraph->getSequence(3)->getLength() == 1);
// original join
SGJoin j1(SGSide(SGPosition(1, 0), true),
SGSide(SGPosition(2, 0), false));
CuAssertTrue(testCase, outGraph->getJoin(&j1) != NULL);
// 1 new join added by fragmentation of sequence 0
SGJoin j2(SGSide(SGPosition(0, 4), false),
SGSide(SGPosition(1, 0), true));
CuAssertTrue(testCase, outGraph->getJoin(&j2) != NULL);
// 1 new join added by fragmentation of sequence 1
SGJoin j3(SGSide(SGPosition(2, 0), false),
SGSide(SGPosition(3, 0), true));
CuAssertTrue(testCase, outGraph->getJoin(&j3) != NULL);
CuAssertTrue(testCase, outGraph->getJoinSet()->size() == 3);
}
///////////////////////////////////////////////////////////
// doubleCut test
// case that first came up in camel-brca1 where segment cut
// failed because of equivalent join sides being processed
// when fragmenting segment
//
///////////////////////////////////////////////////////////
void doubleCutTest(CuTest *testCase)
{
CuAssertTrue(testCase, true);
SideGraph sg;
SGPosition p1a(0, 5);
SGPosition p1b(0, 6);
SGPosition p1c(0, 10);
sg.addSequence(new SGSequence(0, 20, "seq1"));
sg.addJoin(new SGJoin(SGSide(p1a, false), SGSide(p1c, true)));
sg.addJoin(new SGJoin(SGSide(p1b, true), SGSide(p1c, false)));
vector<string> bases(1);
bases[0] = "ACCTGACCATAGGCATGGGC";
vector<SGNamedPath> paths;
Side2Seq converter;
converter.init(&sg, &bases, &paths);
try
{
converter.convert();
}
catch(exception& e)
{
cerr << "Exception caught " << e.what() << endl;
CuAssertTrue(testCase, false);
}
const SideGraph* outGraph = converter.getOutGraph();
CuAssertTrue(testCase, outGraph->getNumSequences() == 4);
CuAssertTrue(testCase, outGraph->getSequence(0)->getLength() == 6);
CuAssertTrue(testCase, outGraph->getSequence(1)->getLength() == 4);
CuAssertTrue(testCase, outGraph->getSequence(2)->getLength() == 1);
CuAssertTrue(testCase, outGraph->getSequence(3)->getLength() == 9);
}
int main(int argc, char** argv)
{
if (argc < 2)
{
help(argv);
return 1;
}
int pageSize = SGClient::DefaultPageSize;
bool upperCase = false;
bool seqPaths = false;
bool skipPaths = false;
optind = 1;
while (true)
{
static struct option long_options[] =
{
{"help", no_argument, 0, 'h'},
{"page", required_argument, 0, 'p'},
{"upper", no_argument, 0, 'u'},
{"paths", no_argument, 0, 'a'},
{"no-paths", no_argument, 0, 'n'}
};
int option_index = 0;
int c = getopt_long(argc, argv, "hp:uan", long_options, &option_index);
if (c == -1)
{
break;
}
switch(c)
{
case 'h':
case '?':
help(argv);
exit(1);
case 'p':
pageSize = atoi(optarg);
break;
case 'u':
upperCase = true;
break;
case 'a':
seqPaths = true;
break;
case 'n':
skipPaths = true;
break;
default:
abort();
}
}
Download::init();
string url = argv[optind];
SGClient sgClient;
sgClient.setURL(url);
sgClient.setOS(&cerr);
sgClient.setPageSize(pageSize);
sgClient.setSkipPaths(skipPaths);
// ith element is bases for sequence with id i in side graph
vector<string> bases;
// ith element is <name, segment vector> for allele i
vector<SGClient::NamedPath> paths;
const SideGraph* sg = sgClient.downloadGraph(bases, paths);
// convert side graph into sequence graph (which is stored
cerr << "Converting Side Graph to VG Sequence Graph" << endl;
Side2Seq converter;
converter.init(sg, &bases, &paths, upperCase, seqPaths, "&SG_");
converter.convert();
const SideGraph* outGraph = converter.getOutGraph();
const vector<string>& outBases = converter.getOutBases();
const vector<SGClient::NamedPath>& outPaths = converter.getOutPaths();
// write to vg json
cerr << "Writing VG JSON to stdout" << endl;
SG2VGJSON jsonWriter;
jsonWriter.init(&cout);
jsonWriter.writeGraph(outGraph, outBases, outPaths);
/*
cerr << "INPUT " << endl;
cerr << *sgClient.getSideGraph() << endl;
cerr << "OUTPUT " << endl;
cerr << *outGraph << endl;
for (int i = 0; i < outPaths.size(); ++i)
{
cerr << "path " << i << ":\n";
cerr << "input " << "name=" << outPaths[i].first;
for (int j = 0; j < outPaths[i].second.size(); ++j)
{
cerr << outPaths[i].second[j] << ", ";
}
cerr << endl;
cerr << "output " << "name=" << outPaths[i].first;
for (int j = 0; j < outPaths[i].second.size(); ++j)
{
cerr << outPaths[i].second[j] << ", ";
}
cerr << endl;
}
*/
Download::cleanup();
}
void chopTest(CuTest *testCase)
{
CuAssertTrue(testCase, true);
// build easy side graph with one snp.
SideGraph sg;
sg.addSequence(new SGSequence(0, 10, "seq1"));
sg.addSequence(new SGSequence(1, 1, "snp"));
sg.addJoin(new SGJoin(SGSide(SGPosition(0, 3), false),
SGSide(SGPosition(1, 0), true)));
sg.addJoin(new SGJoin(SGSide(SGPosition(1, 0), false),
SGSide(SGPosition(0, 5), true)));
vector<string> bases(2);
bases[0] = string(10, 'A');
bases[1] = string(1, 'G');
vector<SGNamedPath> paths(2);
paths[0].first = "path1";
paths[0].second.push_back(SGSegment(SGSide(SGPosition(0, 0), true), 10));
paths[1].first = "path2";
paths[1].second.push_back(SGSegment(SGSide(SGPosition(0, 0), true), 4));
paths[1].second.push_back(SGSegment(SGSide(SGPosition(1, 0), true), 1));
paths[1].second.push_back(SGSegment(SGSide(SGPosition(0, 5), true), 5));
Side2Seq converter;
converter.init(&sg, &bases, &paths, false, false, "", 2);
converter.convert();
const SideGraph* outGraph = converter.getOutGraph();
const vector<string> outBases = converter.getOutBases();
const vector<SGNamedPath> outPaths = converter.getOutPaths();
// expect sequences of length 2, 2, 1, 1, 2, 2, 1
CuAssertTrue(testCase, outGraph->getNumSequences() == 7);
CuAssertTrue(testCase, outGraph->getSequence(0)->getLength() == 2);
CuAssertTrue(testCase, outGraph->getSequence(1)->getLength() == 2);
CuAssertTrue(testCase, outGraph->getSequence(2)->getLength() == 1);
CuAssertTrue(testCase, outGraph->getSequence(3)->getLength() == 1);
CuAssertTrue(testCase, outGraph->getSequence(4)->getLength() == 2);
CuAssertTrue(testCase, outGraph->getSequence(5)->getLength() == 2);
CuAssertTrue(testCase, outGraph->getSequence(6)->getLength() == 1);
// expect 5 extra joins
CuAssertTrue(testCase, outGraph->getJoinSet()->size() == 7);
SGJoin j1(SGSide(SGPosition(0, 1), false),
SGSide(SGPosition(1, 0), true));
CuAssertTrue(testCase, outGraph->getJoin(&j1) != NULL);
SGJoin j2(SGSide(SGPosition(1, 1), false),
SGSide(SGPosition(2, 0), true));
CuAssertTrue(testCase, outGraph->getJoin(&j2) != NULL);
SGJoin j3(SGSide(SGPosition(1, 1), false),
SGSide(SGPosition(6, 0), true));
CuAssertTrue(testCase, outGraph->getJoin(&j3) != NULL);
SGJoin j4(SGSide(SGPosition(2, 0), false),
SGSide(SGPosition(3, 0), true));
CuAssertTrue(testCase, outGraph->getJoin(&j4) != NULL);
SGJoin j5(SGSide(SGPosition(6, 0), false),
SGSide(SGPosition(3, 0), true));
CuAssertTrue(testCase, outGraph->getJoin(&j5) != NULL);
SGJoin j6(SGSide(SGPosition(3, 0), false),
SGSide(SGPosition(4, 0), true));
CuAssertTrue(testCase, outGraph->getJoin(&j6) != NULL);
SGJoin j7(SGSide(SGPosition(4, 1), false),
SGSide(SGPosition(5, 0), true));
CuAssertTrue(testCase, outGraph->getJoin(&j7) != NULL);
// both paths should be broken up over the six segments
CuAssertTrue(testCase, outPaths.size() == 2);
CuAssertTrue(testCase, outPaths[0].second.size() == 6);
CuAssertTrue(testCase, outPaths[0].second.size() == 6);
}
///////////////////////////////////////////////////////////
// Make sure trivial graph runs through ok
//
///////////////////////////////////////////////////////////
void simpleTest1(CuTest *testCase)
{
CuAssertTrue(testCase, true);
// build easy side graph with one snp.
SideGraph sg;
sg.addSequence(new SGSequence(0, 10, "seq1"));
sg.addSequence(new SGSequence(1, 1, "snp"));
sg.addJoin(new SGJoin(SGSide(SGPosition(0, 3), false),
SGSide(SGPosition(1, 0), true)));
sg.addJoin(new SGJoin(SGSide(SGPosition(1, 0), false),
SGSide(SGPosition(0, 5), true)));
vector<string> bases(2);
bases[0] = string(10, 'A');
bases[1] = string(1, 'G');
vector<SGNamedPath> paths(2);
paths[0].first = "path1";
paths[0].second.push_back(SGSegment(SGSide(SGPosition(0, 0), true), 10));
paths[1].first = "path2";
paths[1].second.push_back(SGSegment(SGSide(SGPosition(0, 0), true), 4));
paths[1].second.push_back(SGSegment(SGSide(SGPosition(1, 0), true), 1));
paths[1].second.push_back(SGSegment(SGSide(SGPosition(0, 5), true), 5));
Side2Seq converter;
converter.init(&sg, &bases, &paths);
converter.convert();
const SideGraph* outGraph = converter.getOutGraph();
const vector<string> outBases = converter.getOutBases();
const vector<SGNamedPath> outPaths = converter.getOutPaths();
// expect sequences of length 4, 1, 5, 1
CuAssertTrue(testCase, outGraph->getNumSequences() == 4);
CuAssertTrue(testCase, outGraph->getSequence(0)->getLength() == 4);
CuAssertTrue(testCase, outGraph->getSequence(1)->getLength() == 1);
CuAssertTrue(testCase, outGraph->getSequence(2)->getLength() == 5);
CuAssertTrue(testCase, outGraph->getSequence(3)->getLength() == 1);
// expect 2 extra joins
CuAssertTrue(testCase, outGraph->getJoinSet()->size() == 4);
SGJoin j1(SGSide(SGPosition(0, 3), false),
SGSide(SGPosition(1, 0), true));
CuAssertTrue(testCase, outGraph->getJoin(&j1) != NULL);
SGJoin j2(SGSide(SGPosition(0, 3), false),
SGSide(SGPosition(3, 0), true));
CuAssertTrue(testCase, outGraph->getJoin(&j2) != NULL);
SGJoin j3(SGSide(SGPosition(1, 0), false),
SGSide(SGPosition(2, 0), true));
CuAssertTrue(testCase, outGraph->getJoin(&j3) != NULL);
SGJoin j4(SGSide(SGPosition(3, 0), false),
SGSide(SGPosition(2, 0), true));
CuAssertTrue(testCase, outGraph->getJoin(&j4) != NULL);
for (int i = 0; i < outBases.size(); ++i)
{
CuAssertTrue(testCase, outBases[i].length() ==
outGraph->getSequence(i)->getLength());
if (i == 3)
{
CuAssertTrue(testCase, outBases[i] == bases[1]);
}
else
{
CuAssertTrue(testCase, outBases[i] == string(outBases[i].length(), 'A'));
}
}
// both paths should be broken up over the three segments
CuAssertTrue(testCase, outPaths.size() == 2);
CuAssertTrue(testCase, outPaths[0].second.size() == 3);
CuAssertTrue(testCase, outPaths[0].second.size() == 3);
CuAssertTrue(testCase, outPaths[0].second[0] ==
SGSegment(SGSide(SGPosition(0, 0), true), 4));
CuAssertTrue(testCase, outPaths[0].second[1] ==
SGSegment(SGSide(SGPosition(1, 0), true), 1));
CuAssertTrue(testCase, outPaths[0].second[2] ==
SGSegment(SGSide(SGPosition(2, 0), true), 5));
CuAssertTrue(testCase, outPaths[1].second[0] ==
SGSegment(SGSide(SGPosition(0, 0), true), 4));
CuAssertTrue(testCase, outPaths[1].second[1] ==
SGSegment(SGSide(SGPosition(3, 0), true), 1));
CuAssertTrue(testCase, outPaths[1].second[2] ==
SGSegment(SGSide(SGPosition(2, 0), true), 5));
CuAssertTrue(testCase, outPaths[0].first == paths[0].first);
CuAssertTrue(testCase, outPaths[1].first == paths[1].first);
}
///////////////////////////////////////////////////////////
// inversion test
// Make sure all 4 types of joins are converted
//
///////////////////////////////////////////////////////////
void inversionTest(CuTest *testCase)
{
CuAssertTrue(testCase, true);
SideGraph sg;
sg.addSequence(new SGSequence(0, 20, "seq1"));
sg.addSequence(new SGSequence(1, 10, "seq2"));
SGPosition p1a(0, 5);
SGPosition p1b(0, 9);
SGPosition p1c(0, 10);
SGPosition p2a(1, 2);
SGPosition p2b(1, 4);
SGPosition p2c(1, 6);
// R->F
sg.addJoin(new SGJoin(SGSide(p1a, false), SGSide(p2a, true)));
// R->R
sg.addJoin(new SGJoin(SGSide(p2a, false), SGSide(p2b, false)));
// F->R
sg.addJoin(new SGJoin(SGSide(p2b, true), SGSide(p1c, false)));
// F->F
sg.addJoin(new SGJoin(SGSide(p1b, true), SGSide(p2c, true)));
vector<string> bases(2);
bases[0] = "ACCTGACCATAGGCATGGGC";
bases[1] = "TCCGCCTAAA";
vector<SGNamedPath> paths(2);
// follow joins "left to right"
paths[0].first = "path1";
paths[0].second.push_back(SGSegment(SGSide(SGPosition(0, 0), true), 6));
paths[0].second.push_back(SGSegment(SGSide(p2a, true), 1));
paths[0].second.push_back(SGSegment(SGSide(p2b, false), 1));
paths[0].second.push_back(SGSegment(SGSide(p1b, false), 2));
paths[0].second.push_back(SGSegment(SGSide(p2c, true), 2));
// same path but in other direction
paths[1].first = "path2";
paths[1].second.push_back(SGSegment(SGSide(SGPosition(1, 9), false), 2));
paths[1].second.push_back(SGSegment(SGSide(p1b, true), 2));
paths[1].second.push_back(SGSegment(SGSide(p2b, true), 1));
paths[1].second.push_back(SGSegment(SGSide(p2a, false), 1));
paths[1].second.push_back(SGSegment(SGSide(p1a, false), 6));
Side2Seq converter;
converter.init(&sg, &bases, &paths);
try
{
converter.convert();
}
catch(exception& e)
{
// inside convert, are some sanity checks making sure the paths
// get converted properly. we'll be lazy and rely on this code
// to make sure the paths got converted.
cerr << "Exception caught " << e.what() << endl;
CuAssertTrue(testCase, false);
}
const SideGraph* outGraph = converter.getOutGraph();
const vector<string> outBases = converter.getOutBases();
const vector<SGNamedPath> outPaths = converter.getOutPaths();
// first sequence should be cut up into 4, and 2nd into 7
CuAssertTrue(testCase, outGraph->getNumSequences() == 11);
CuAssertTrue(testCase, outGraph->getSequence(0)->getLength() == 6);
CuAssertTrue(testCase, outGraph->getSequence(1)->getLength() == 3);
CuAssertTrue(testCase, outGraph->getSequence(2)->getLength() == 2);
CuAssertTrue(testCase, outGraph->getSequence(3)->getLength() == 9);
CuAssertTrue(testCase, outGraph->getSequence(4)->getLength() == 2);
CuAssertTrue(testCase, outGraph->getSequence(5)->getLength() == 1);
CuAssertTrue(testCase, outGraph->getSequence(6)->getLength() == 1);
CuAssertTrue(testCase, outGraph->getSequence(7)->getLength() == 1);
CuAssertTrue(testCase, outGraph->getSequence(8)->getLength() == 1);
CuAssertTrue(testCase, outGraph->getSequence(9)->getLength() == 2);
CuAssertTrue(testCase, outGraph->getSequence(10)->getLength() == 2);
// joins
// 4 original joins as mapped to new graph
SGJoin j1(SGSide(SGPosition(0, 5), false),
SGSide(SGPosition(5, 0), true));
CuAssertTrue(testCase, outGraph->getJoin(&j1) != NULL);
SGJoin j2(SGSide(SGPosition(5, 0), false),
SGSide(SGPosition(7, 0), false));
CuAssertTrue(testCase, outGraph->getJoin(&j2) != NULL);
SGJoin j3(SGSide(SGPosition(7, 0), true),
SGSide(SGPosition(2, 1), false));
CuAssertTrue(testCase, outGraph->getJoin(&j3) != NULL);
SGJoin j4(SGSide(SGPosition(2, 0), true),
SGSide(SGPosition(9, 0), true));
CuAssertTrue(testCase, outGraph->getJoin(&j3) != NULL);
// 3 new joins added by fragmentation to sequence 0
for (int i = 0; i < 3; ++i)
{
SGJoin j5(SGSide(SGPosition(i, outGraph->getSequence(i)->getLength() - 1),
false),
SGSide(SGPosition(i + 1, 0), true));
CuAssertTrue(testCase, outGraph->getJoin(&j5) != NULL);
}
// 5 new joins added by fragmentation to sequence 1
for (int i = 5; i < 9; ++i)
{
SGJoin j5(SGSide(SGPosition(i, outGraph->getSequence(i)->getLength() - 1),
false),
SGSide(SGPosition(i + 1, 0), true));
CuAssertTrue(testCase, outGraph->getJoin(&j5) != NULL);
}
CuAssertTrue(testCase, outGraph->getJoinSet()->size() == 13);
}