Exemplo n.º 1
0
int main(int argc, char **argv)
{
	ReadSet *sequences = NULL;
	RoadMapArray *rdmaps;
	PreGraph *preGraph;
	Graph *graph;
	char *directory, *graphFilename, *preGraphFilename, *seqFilename,
	    *roadmapFilename;
	double coverageCutoff = -1;
	double maxCoverageCutoff = -1;
	double expectedCoverage = -1;
	int longMultCutoff = -1;
	Coordinate minContigLength = -1;
	Coordinate minContigKmerLength;
	boolean *dubious = NULL;
	Coordinate insertLength[CATEGORIES];
	Coordinate insertLengthLong = -1;
	Coordinate std_dev[CATEGORIES];
	Coordinate std_dev_long = -1;
	short int accelerationBits = 24;
	boolean readTracking = false;
	boolean exportAssembly = false;
	boolean unusedReads = false;
	boolean estimateCoverage = false;
	boolean estimateCutoff = false;
	FILE *file;
	int arg_index, arg_int;
	double arg_double;
	char *arg;
	Coordinate *sequenceLengths = NULL;
	Category cat;
	boolean scaffolding = true;
	int pebbleRounds = 1;
	long long longlong_var;
	short int short_var;

	setProgramName("velvetg");

	for (cat = 0; cat < CATEGORIES; cat++) {
		insertLength[cat] = -1;
		std_dev[cat] = -1;
	}

	// Error message
	if (argc == 1) {
		puts("velvetg - de Bruijn graph construction, error removal and repeat resolution");
		printf("Version %i.%i.%2.2i\n", VERSION_NUMBER,
		       RELEASE_NUMBER, UPDATE_NUMBER);
		puts("\nCopyright 2007, 2008 Daniel Zerbino ([email protected])");
		puts("This is free software; see the source for copying conditions.  There is NO");
		puts("warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.\n");
		puts("Compilation settings:");
		printf("CATEGORIES = %i\n", CATEGORIES);
		printf("MAXKMERLENGTH = %i\n", MAXKMERLENGTH);
		puts("");
		printUsage();
		return 1;
	}

	if (strcmp(argv[1], "--help") == 0) {
		printUsage();
		return 0;
	}

	// Memory allocation 
	directory = argv[1];
	graphFilename = mallocOrExit(strlen(directory) + 100, char);
	preGraphFilename =
	    mallocOrExit(strlen(directory) + 100, char);
	roadmapFilename = mallocOrExit(strlen(directory) + 100, char);
	seqFilename = mallocOrExit(strlen(directory) + 100, char);
	// Argument parsing
	for (arg_index = 2; arg_index < argc; arg_index++) {
		arg = argv[arg_index++];
		if (arg_index >= argc) {
			puts("Unusual number of arguments!");
			printUsage();
			exit(1);
		}

		if (strcmp(arg, "-cov_cutoff") == 0) {
			if (strcmp(argv[arg_index], "auto") == 0) {
				estimateCutoff = true;
			} else {
				sscanf(argv[arg_index], "%lf", &coverageCutoff);
			}
		} else if (strcmp(arg, "-exp_cov") == 0) {
			if (strcmp(argv[arg_index], "auto") == 0) {
				estimateCoverage = true;
				readTracking = true;
			} else {
				sscanf(argv[arg_index], "%lf", &expectedCoverage);
				if (expectedCoverage > 0)
					readTracking = true;
			}
		} else if (strcmp(arg, "-ins_length") == 0) {
			sscanf(argv[arg_index], "%lli", &longlong_var);
			insertLength[0] = (Coordinate) longlong_var;
			if (insertLength[0] < 0) {
				printf("Invalid insert length: %lli\n",
				       (long long) insertLength[0]);
				exit(1);
			}
		} else if (strcmp(arg, "-ins_length_sd") == 0) {
			sscanf(argv[arg_index], "%lli", &longlong_var);
			std_dev[0] = (Coordinate) longlong_var;
			if (std_dev[0] < 0) {
				printf("Invalid std deviation: %lli\n",
				       (long long) std_dev[0]);
				exit(1);
			}
		} else if (strcmp(arg, "-ins_length_long") == 0) {
			sscanf(argv[arg_index], "%lli", &longlong_var);
			insertLengthLong = (Coordinate) longlong_var;
		} else if (strcmp(arg, "-ins_length_long_sd") == 0) {
			sscanf(argv[arg_index], "%lli", &longlong_var);
			std_dev_long = (Coordinate) longlong_var;
		} else if (strncmp(arg, "-ins_length", 11) == 0
			   && strchr(arg, 'd') == NULL) {
			sscanf(arg, "-ins_length%hi", &short_var);
			cat = (Category) short_var;
			if (cat < 1 || cat > CATEGORIES) {
				printf("Unknown option: %s\n", arg);
				exit(1);
			}
			sscanf(argv[arg_index], "%lli", &longlong_var);
			insertLength[cat - 1] = (Coordinate) longlong_var;
			if (insertLength[cat - 1] < 0) {
				printf("Invalid insert length: %lli\n",
				       (long long) insertLength[cat - 1]);
				exit(1);
			}
		} else if (strncmp(arg, "-ins_length", 11) == 0) {
			sscanf(arg, "-ins_length%hi_sd", &short_var);
			cat = (Category) short_var;
			if (cat < 1 || cat > CATEGORIES) {
				printf("Unknown option: %s\n", arg);
				exit(1);
			}
			sscanf(argv[arg_index], "%lli", &longlong_var);
			std_dev[cat - 1] = (Coordinate) longlong_var;
			if (std_dev[cat - 1] < 0) {
				printf("Invalid std deviation: %lli\n",
				       (long long) std_dev[cat - 1]);
				exit(1);
			}
		} else if (strcmp(arg, "-read_trkg") == 0) {
			readTracking =
			    (strcmp(argv[arg_index], "yes") == 0);
		} else if (strcmp(arg, "-scaffolding") == 0) {
			scaffolding =
			    (strcmp(argv[arg_index], "yes") == 0);
		} else if (strcmp(arg, "-amos_file") == 0) {
			exportAssembly =
			    (strcmp(argv[arg_index], "yes") == 0);
		} else if (strcmp(arg, "-min_contig_lgth") == 0) {
			sscanf(argv[arg_index], "%lli", &longlong_var);
			minContigLength = (Coordinate) longlong_var;
		} else if (strcmp(arg, "-accel_bits") == 0) {
			sscanf(argv[arg_index], "%hi", &accelerationBits);
			if (accelerationBits < 0) {
				printf
				    ("Illegal acceleration parameter: %s\n",
				     argv[arg_index]);
				printUsage();
				return -1;
			}
		} else if (strcmp(arg, "-max_branch_length") == 0) {
			sscanf(argv[arg_index], "%i", &arg_int);
			setMaxReadLength(arg_int);
			setLocalMaxReadLength(arg_int);
		} else if (strcmp(arg, "-max_divergence") == 0) {
			sscanf(argv[arg_index], "%lf", &arg_double);
			setMaxDivergence(arg_double);
			setLocalMaxDivergence(arg_double);
		} else if (strcmp(arg, "-max_gap_count") == 0) {
			sscanf(argv[arg_index], "%i", &arg_int);
			setMaxGaps(arg_int);
			setLocalMaxGaps(arg_int);
		} else if (strcmp(arg, "-min_pair_count") == 0) {
			sscanf(argv[arg_index], "%i", &arg_int);
			setUnreliableConnectionCutoff(arg_int);
		} else if (strcmp(arg, "-max_coverage") == 0) {
			sscanf(argv[arg_index], "%lf", &maxCoverageCutoff);
		} else if (strcmp(arg, "-long_mult_cutoff") == 0) {
			sscanf(argv[arg_index], "%i", &longMultCutoff);
			setMultiplicityCutoff(longMultCutoff);
		} else if (strcmp(arg, "-unused_reads") == 0) {
			unusedReads =
			    (strcmp(argv[arg_index], "yes") == 0);
			if (unusedReads)
				readTracking = true;
		} else if (strcmp(arg, "--help") == 0) {
			printUsage();
			return 0;	
		} else {
			printf("Unknown option: %s;\n", arg);
			printUsage();
			return 1;
		}
	}

	// Bookkeeping
	logInstructions(argc, argv, directory);

	strcpy(seqFilename, directory);
	strcat(seqFilename, "/Sequences");

	strcpy(roadmapFilename, directory);
	strcat(roadmapFilename, "/Roadmaps");

	strcpy(preGraphFilename, directory);
	strcat(preGraphFilename, "/PreGraph");

	if (!readTracking) {
		strcpy(graphFilename, directory);
		strcat(graphFilename, "/Graph");
	} else {
		strcpy(graphFilename, directory);
		strcat(graphFilename, "/Graph2");
	}

	// Graph uploading or creation
	if ((file = fopen(graphFilename, "r")) != NULL) {
		fclose(file);
		graph = importGraph(graphFilename);
	} else if ((file = fopen(preGraphFilename, "r")) != NULL) {
		fclose(file);
		sequences = importReadSet(seqFilename);
		convertSequences(sequences);
		graph =
		    importPreGraph(preGraphFilename, sequences,
				   readTracking, accelerationBits);
		sequenceLengths =
		    getSequenceLengths(sequences, getWordLength(graph));
		correctGraph(graph, sequenceLengths);
		exportGraph(graphFilename, graph, sequences->tSequences);
	} else if ((file = fopen(roadmapFilename, "r")) != NULL) {
		fclose(file);
		rdmaps = importRoadMapArray(roadmapFilename);
		preGraph = newPreGraph_pg(rdmaps, seqFilename);
		clipTips_pg(preGraph);
		exportPreGraph_pg(preGraphFilename, preGraph);
		destroyPreGraph_pg(preGraph);

		sequences = importReadSet(seqFilename);
		convertSequences(sequences);
		graph =
		    importPreGraph(preGraphFilename, sequences,
				   readTracking, accelerationBits);
		sequenceLengths =
		    getSequenceLengths(sequences, getWordLength(graph));
		correctGraph(graph, sequenceLengths);
		exportGraph(graphFilename, graph, sequences->tSequences);
	} else {
		puts("No Roadmap file to build upon! Please run velveth (see manual)");
		exit(1);
	}

	// Set insert lengths and their standard deviations
	for (cat = 0; cat < CATEGORIES; cat++) {
		if (insertLength[cat] > -1 && std_dev[cat] < 0)
			std_dev[cat] = insertLength[cat] / 10;
		setInsertLengths(graph, cat,
				 insertLength[cat], std_dev[cat]);
	}

	if (insertLengthLong > -1 && std_dev_long < 0)
		std_dev_long = insertLengthLong / 10;
	setInsertLengths(graph, CATEGORIES,
			 insertLengthLong, std_dev_long);

	// Coverage cutoff
	if (expectedCoverage < 0 && estimateCoverage == true) {
		expectedCoverage = estimated_cov(graph);
		if (coverageCutoff < 0) {
			coverageCutoff = expectedCoverage / 2;
			estimateCutoff = true;
		}
	} else { 
		estimateCoverage = false;
		if (coverageCutoff < 0 && estimateCutoff) 
			coverageCutoff = estimated_cov(graph) / 2;
		else 
			estimateCutoff = false;
	}

	if (coverageCutoff < 0) {
		puts("WARNING: NO COVERAGE CUTOFF PROVIDED");
		puts("Velvet will probably leave behind many detectable errors");
		puts("See manual for instructions on how to set the coverage cutoff parameter");
	}

	dubious =
	    removeLowCoverageNodesAndDenounceDubiousReads(graph,
							  coverageCutoff);
	removeHighCoverageNodes(graph, maxCoverageCutoff);
	clipTipsHard(graph);

	if (expectedCoverage > 0) {
		if (sequences == NULL) {
			sequences = importReadSet(seqFilename);
			convertSequences(sequences);
		}

		// Mixed length sequencing
		readCoherentGraph(graph, isUniqueSolexa, expectedCoverage,
				  sequences);

		// Paired ends module
		createReadPairingArray(sequences);
		for (cat = 0; cat < CATEGORIES; cat++) 
			if(pairUpReads(sequences, 2 * cat + 1))
				pebbleRounds++;

		if (pairUpReads(sequences, 2 * CATEGORIES + 1))
			pebbleRounds++;

		detachDubiousReads(sequences, dubious);
		activateGapMarkers(graph);
		for ( ;pebbleRounds > 0; pebbleRounds--)
			exploitShortReadPairs(graph, sequences, dubious, scaffolding);
	} else {
		puts("WARNING: NO EXPECTED COVERAGE PROVIDED");
		puts("Velvet will be unable to resolve any repeats");
		puts("See manual for instructions on how to set the expected coverage parameter");
	}

	free(dubious);

	concatenateGraph(graph);

	if (minContigLength < 2 * getWordLength(graph))
		minContigKmerLength = getWordLength(graph);
	else
		minContigKmerLength = minContigLength - getWordLength(graph) + 1;		

	strcpy(graphFilename, directory);
	strcat(graphFilename, "/contigs.fa");
	exportLongNodeSequences(graphFilename, graph, minContigKmerLength); 

	strcpy(graphFilename, directory);
	strcat(graphFilename, "/stats.txt");
	displayGeneralStatistics(graph, graphFilename);

	if (sequences == NULL) {
		sequences = importReadSet(seqFilename);
		convertSequences(sequences);
	}

	strcpy(graphFilename, directory);
	strcat(graphFilename, "/LastGraph");
	exportGraph(graphFilename, graph, sequences->tSequences);

	if (exportAssembly) {
		strcpy(graphFilename, directory);
		strcat(graphFilename, "/velvet_asm.afg");
		exportAMOSContigs(graphFilename, graph, minContigKmerLength, sequences);
	}

	if (unusedReads)
		exportUnusedReads(graph, sequences, minContigKmerLength, directory);

	if (estimateCoverage) 
		printf("Estimated Coverage = %f\n", expectedCoverage);
	if (estimateCutoff) 
		printf("Estimated Coverage cutoff = %f\n", coverageCutoff);

	logFinalStats(graph, minContigKmerLength, directory);

	destroyGraph(graph);
	free(graphFilename);
	free(preGraphFilename);
	free(seqFilename);
	free(roadmapFilename);
	destroyReadSet(sequences);
	return 0;
}
Exemplo n.º 2
0
int main(int argc, char **argv)
{
	ReadSet *allSequences = NULL;
	SplayTable *splayTable;
	int hashLength, hashLengthStep, hashLengthMax, h;
	char *directory, *filename, *seqFilename, *baseSeqName, *buf;
	char * token;
	boolean double_strand = true;
	boolean noHash = false;
	boolean multiple_kmers = false;
	char buffer[100];
	DIR *dir;

	setProgramName("velveth");

	if (argc < 4) {
		printf("velveth - simple hashing program\n");
		printf("Version %i.%i.%2.2i\n", VERSION_NUMBER,
		       RELEASE_NUMBER, UPDATE_NUMBER);
		printf("\nCopyright 2007, 2008 Daniel Zerbino ([email protected])\n");
		printf("This is free software; see the source for copying conditions.  There is NO\n");
		printf("warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.\n\n");
		printf("Compilation settings:\n");
		printf("CATEGORIES = %i\n", CATEGORIES);
		printf("MAXKMERLENGTH = %i\n", MAXKMERLENGTH);
#ifdef _OPENMP
		puts("OPENMP");
#endif
#ifdef LONGSEQUENCES
		puts("LONGSEQUENCES");
#endif
#ifdef BIGASSEMBLY
		puts("BIGASSEMBLY");
#endif
#ifdef COLOR
		puts("COLOR");
#endif
#ifdef DEBUG
		puts("DEBUG");
#endif
		printf("\n");
		printUsage();
		return 0;
	}

	strcpy(buffer, argv[2]);
	token = strtok(buffer, ",");
	hashLength = atoi(token);
	token = strtok(NULL, ",");
	if (token == NULL) {
		multiple_kmers = false;
		hashLengthMax = hashLength + 1;
	} else {
		multiple_kmers = true;
		hashLengthMax = atoi(token);
	}
	token = strtok(NULL, ",");
	if (token == NULL) {
		hashLengthStep = 2;
	} else {
		hashLengthStep = atoi(token);
	}

	if (hashLength > MAXKMERLENGTH) {
		velvetLog
		    ("Velvet can't handle k-mers as long as %i! We'll stick to %i if you don't mind.\n",
		     hashLength, MAXKMERLENGTH);
		hashLength = MAXKMERLENGTH;
	} 
	if (hashLength <= 0) {
		velvetLog("Invalid hash length: %s\n", argv[2]);
		printUsage();
		return 0;
	} 
	if (hashLength % 2 == 0) {
		velvetLog
		    ("Velvet can't work with even length k-mers, such as %i. We'll use %i instead, if you don't mind.\n",
		     hashLength, hashLength - 1);
		hashLength--;
	} 

	if (multiple_kmers) {
		if (hashLengthMax > MAXKMERLENGTH + 1) {
			velvetLog
			    ("Velvet can't handle k-mers as long as %i! We'll stick to %i if you don't mind.\n",
			     hashLengthMax, MAXKMERLENGTH + 1);
			hashLengthMax = MAXKMERLENGTH + 1;
		} 
		if (hashLengthMax <= hashLength) {
			velvetLog("hashLengthMin < hashLengthMax is required %s", argv[2]);
			printUsage();
			return 0;
		} 

		if (hashLengthStep <= 0) {
			velvetLog("Non-positive hash length! Setting it to 2\n");
			hashLengthStep = 2;
		}
		if (hashLengthStep % 2 == 1) {
			velvetLog
			    ("Velvet can't work with an odd length k-mer step, such as %i. We'll use %i instead, if you don't mind.\n",
			     hashLengthStep, hashLengthStep + 1);
			hashLengthStep++;
		}
	}

	// check if binary sequences should be used
	int argIndex;
	for (argIndex = 3; argIndex < argc; argIndex++)
		if (strcmp(argv[argIndex], "-create_binary") == 0 || strcmp(argv[argIndex], "-reuse_binary") == 0)
			setCreateBinary(true);

	for (h = hashLength; h < hashLengthMax; h += hashLengthStep) {

		resetWordFilter(h);

		buf = mallocOrExit(2 * strlen(argv[1]) + 500, char);

		if ( multiple_kmers ) {
			sprintf(buf,"%s_%d",argv[1],h);
			directory = mallocOrExit(strlen(buf) + 100, char);
			strcpy(directory,buf);
		} else 
			directory = argv[1];

		filename = mallocOrExit(strlen(directory) + 100, char);
		seqFilename = mallocOrExit(strlen(directory) + 100, char);
		baseSeqName = mallocOrExit(100, char);

		dir = opendir(directory);

		if (dir == NULL)
			mkdir(directory, 0777);
		else {
			sprintf(buf, "%s/PreGraph", directory);
			remove(buf);
			sprintf(buf, "%s/Graph", directory);
			remove(buf);
			sprintf(buf, "%s/Graph2", directory);
			remove(buf);
			sprintf(buf, "%s/Graph3", directory);
			remove(buf);
			sprintf(buf, "%s/Graph4", directory);
			remove(buf);
		}

		logInstructions(argc, argv, directory);

		strcpy(seqFilename, directory);
		if (isCreateBinary()) {
			// use the CNY unified seq writer
			strcpy(baseSeqName, "/CnyUnifiedSeq");
			// remove other style sequences file
			sprintf(buf, "%s/Sequences", directory);
			remove(buf);
		} else {
			strcpy(baseSeqName, "/Sequences");
			// remove other style sequences file
			sprintf(buf, "%s/CnyUnifiedSeq", directory);
			remove(buf);
			sprintf(buf, "%s/CnyUnifiedSeq.names", directory);
			remove(buf);
		}
		strcat(seqFilename, baseSeqName);

		if ( h == hashLength ) {
			parseDataAndReadFiles(seqFilename, argc - 2, &(argv[2]), &double_strand, &noHash);
		} else {
			sprintf(buf,"rm -f %s",seqFilename);
			if (system(buf)) {
				velvetLog("Command failed!\n");
				velvetLog("%s\n", buf);
#ifdef DEBUG
				abort();
#endif
				exit(1);
			}
			if (argv[1][0] == '/')
				sprintf(buf,"ln -s %s_%d%s %s",argv[1],hashLength,baseSeqName,seqFilename);
			else
				sprintf(buf,"ln -s `pwd`/%s_%d%s %s",argv[1],hashLength,baseSeqName,seqFilename);
			if (system(buf)) {
				velvetLog("Command failed!\n");
				velvetLog("%s\n", buf);
#ifdef DEBUG
				abort();
#endif
				exit(1);
			}
		}

		if (noHash)
			continue;

		splayTable = newSplayTable(h, double_strand);
		if (isCreateBinary()) {
			allSequences = importCnyReadSet(seqFilename);
		} else {
			allSequences = importReadSet(seqFilename);
		}
		velvetLog("%li sequences in total.\n", (long) allSequences->readCount);

		strcpy(filename, directory);
		strcat(filename, "/Roadmaps");
		inputSequenceArrayIntoSplayTableAndArchive(allSequences,
							   splayTable, filename, seqFilename);

		destroySplayTable(splayTable);
		if (dir)
			closedir(dir);
		if (directory != argv[1])
			free(directory);
		free(filename);
		free(seqFilename);
		free(baseSeqName);
		free(buf);
		if (allSequences) {
			destroyReadSet(allSequences);
		}
	}