My complete changes on my laptop, with specificity bug fix + ahocorasick + sets

git-svn-id: https://www.grenoble.prabi.fr/svn/LECASofts/ecoPrimers/trunk@393 60f365c0-8329-0410-b2a4-ec073aeeaa1d

src/global.mk

@@ -3,7 +3,8 @@ LIBPATH= -Llibapat -LlibecoPCR -Llibecoprimer -Llibthermo
 MAKEDEPEND = gcc -D$(MACHINE) -M $(CPPFLAGS) -o $*.d $<
 
 CC=gcc
-CFLAGS= -W -Wall -O5 -m64 
+CFLAGS= -W -Wall -m64 -g
+#CFLAGS= -W -Wall -O5 -m64 -g
 #CFLAGS= -W -Wall -O0 -m64  -g
 #CFLAGS= -W -Wall -O5 -fast -g
 

src/libecoprimer/Makefile

@@ -15,7 +15,8 @@ SOURCES = goodtaxon.c \
           taxstats.c \
           apat_search.c \
 		  filtering.c \
-		  PrimerSets.c
+		  PrimerSets.c \
+		  ahocorasick.c
 
 SRCS=$(SOURCES)
          

src/libecoprimer/PrimerSets.h

@@ -13,6 +13,8 @@ typedef struct {
 	float set_lmean;
 	float set_lcov;
 	float set_score;
+	int32_t set_intaxa;
+	int32_t set_wi_cnt;
 }pairset;
 
 typedef struct{
@@ -33,9 +35,24 @@ typedef struct{
 void add_pair_in_set (pairset *pair_set, int32_t pset_idx, int32_t prb_idx, SetParams *pparams);
 void get_next_pair_options (int *pair_wi_count_sorted_ids, pairset *pair_set, SetParams *pparams);
 float get_links_distribution (int prb_idx, pairset *prob_set, SetParams *pparams);
-pairset build_primers_set (ppair_t* sortedpairs, int32_t sorted_count, pecodnadb_t seqdb,
-		poptions_t options);
+pairset build_primers_set_greedy_spc (SetParams *pparams);
 void get_set_mean_cov_stats (pairset *prob_set, SetParams *pparams);
 void some_other_set_possibilities (pairset *pair_set,
 		ppair_t * sortedpairs, int32_t sorted_count, pecodnadb_t seqdb, poptions_t options);
+void sets_by_SimulatedAnealing (pairset *pair_set,
+		ppair_t * sortedpairs, int32_t sorted_count, pecodnadb_t seqdb, poptions_t options);
+void sets_by_TabuSearch (pairset *pair_set,
+		ppair_t * sortedpairs, int32_t sorted_count, pecodnadb_t seqdb, poptions_t options);
+pairset * sets_by_BruteForce (ppair_t * sortedpairs, 
+		int32_t sorted_count, pecodnadb_t seqdb, poptions_t options);
+pairset * extend_set_randomly (pairset *pair_set, SetParams *params, int extend_to_cnt);
+void build_and_print_sets (ppair_t * sortedpairs, int32_t sorted_count, pecodnadb_t seqdb, poptions_t options);
+int32_t get_next_option_increasing_cov (pairset *pair_set, SetParams *pparams);
+void reset_set_props (pairset *pair_set, SetParams *pparams);
+void primers_graph_graphviz (ppair_t * sortedpairs,
+		int32_t sorted_count, poptions_t options);
+size_t primers_changeSortedArray (ppair_t ** pairs,
+		size_t sorted_count, poptions_t options);
+size_t primers_filterWithGivenLinks (ppair_t ** pairs,
+		size_t sorted_count, poptions_t options);
 #endif 

src/libecoprimer/ahocorasick.h

+/*
+ * ahocorasick.h
+ *
+ *  Created on: 26 march 2011
+ *      Author: tiayyba
+ */
+
+#ifndef H_ahocorasick
+#define H_ahocorasick
+
+#include "ecoprimer.h"
+
+typedef struct aho_output_t{
+	uint32_t wordidx; //index of strict word (dont save the word of 64B)
+	bool_t isdirect; //we need to find both direct and reverse words so we must know which one is it
+}aho_output;
+
+typedef struct aho_output_count_t{
+	uint32_t count;
+	aho_output *out_set;
+}aho_output_count;
+
+typedef struct aho_state_t{
+	int32_t id;
+	struct aho_state_t *next[4]; //for labels A=0,C=1,G=2 and T=3
+	struct aho_state_t *fail;
+	aho_output_count output;
+}aho_state;
+
+typedef struct queue_node_t {
+	aho_state *state_node;
+	struct queue_node_t *next;
+}queue_node;
+
+typedef struct{
+	queue_node *first;
+	queue_node *last;
+}aho_queue;
+
+pprimercount_t ahoc_lookforStrictPrimers (pecodnadb_t database, uint32_t seqdbsize,uint32_t exampleCount,
+		                     pwordcount_t words,poptions_t options);
+#endif /* H_ahocorasick */
+

src/libecoprimer/ecoprimer.h

@@ -176,6 +176,7 @@ typedef struct {
 	int				*wellIdentifiedSeqs; //< an array having elements equla to total seqs
 									 // values are either 0 or 1, if seq is well identified
 									 // its 1 else 0
+	int				*coveredSeqs; //< an array having elements equal to total seqs, 1 if seq is covered else 0
 
 					// these statistics are relative to inexample sequences
 
@@ -291,6 +292,9 @@ typedef struct {
 	PNNParams	  pnparm;
 	bool_t		  print_sets_of_primers;
 	float		  specificity_threshold;
+	int			  links_cnt;
+	float		  max_links_percent;
+	bool_t		  filter_on_links;
 } options_t, *poptions_t;
 
 typedef ecoseq_t  **pecodnadb_t;
@@ -350,7 +354,7 @@ int32_t getrankdbstats(pecodnadb_t seqdb,
 					   uint32_t seqdbsize,
 					   ecotaxonomy_t *taxonomy,
 					   poptions_t options);
-float taxonomycoverage(ppair_t pair, poptions_t options);
+float taxonomycoverage(ppair_t pair, poptions_t options, pecodnadb_t seqdb,uint32_t seqdbsize);
 char ecoComplementChar(char base);
 void taxonomyspecificity (ppair_t pair, pecodnadb_t seqdb,uint32_t seqdbsize);
 

src/libecoprimer/filtering.c

@@ -114,6 +114,8 @@ static	int32_t *ecoFilteringHashSequence(int32_t *dest,
 		error<<= 1;
 		error&=ERRORMASK(FWORDSIZE);
 
+		//code = -1;
+		//if((*base) >= 'A' && (*base) <= 'Z')
 		code = encoder[(*base) - 'A'];
 		if (code <0)
 		{
@@ -154,7 +156,7 @@ int32_t *filteringSeq(pecodnadb_t database, uint32_t seqdbsize,
 
 	for (i=0;i<seqdbsize;i++)
     {
-    	if (database[i]->isexample)
+    	if (database[i]->isexample && database[i]->SQ_length > options->primer_length)
     	{
     		j++;
     		wordscount=ecoFilteringHashSequence(wordscount,

src/libecoprimer/pairs.c

@@ -179,7 +179,7 @@ static void buildPrimerPairsForOneSeq(uint32_t seqid,
 	uint32_t           i,j,k;
 	uint32_t           matchcount=0;
 	pprimermatch_t     matches = NULL;
-	primermatchcount_t seqmatchcount;
+	//primermatchcount_t seqmatchcount;
 	ppair_t            pcurrent;
 	pair_t			   current;
 	pprimer_t		   wswp;
@@ -189,9 +189,9 @@ static void buildPrimerPairsForOneSeq(uint32_t seqid,
 	//char 			   prmr[50];
 	//float			   mtemp;
 	word_t 			   w1, w1a, omask = (0x1L << (options->strict_three_prime*2)) -1;
-	word_t 			   w2, w2a, wtmp;
+	word_t 			   w2, w2a;//, wtmp;
 	uint32_t   		   bp1,bp2;
-	
+		
 	//prmr[options->primer_length] = '\0';
 	
 	for (i=0;i < primers->size; i++)
@@ -252,16 +252,17 @@ static void buildPrimerPairsForOneSeq(uint32_t seqid,
 		{
 			// For all primers matching the sequence
 
-			//for(j=i+1;
-			//       (j<matchcount)
-			//    && ((distance=matches[j].position - matches[i].position - options->primer_length) < options->lmax);
-			//   j++
-			//   )
+			/*for(j=i+1;
+			       (j<matchcount)
+			    && ((distance=matches[j].position - matches[i].position - options->primer_length) < options->lmax);
+			   j++
+			   )//*/
 			for (j=i+1; j<matchcount; j++)
 			{
 				if (matches[j].position - matches[i].position <= options->primer_length) continue;
 				distance = matches[j].position - matches[i].position - options->primer_length;
 				if (distance >= options->lmax) break;
+				
 
 			   // For all not too far primers
 
@@ -269,9 +270,7 @@ static void buildPrimerPairsForOneSeq(uint32_t seqid,
 		    		&& (distance > options->lmin)
 		    		)
 		       {
-
 		    	   // If possible primer pair
-
 		    	   current.p1 = matches[i].primer;
 		    	   current.asdirect1=matches[i].strand;
 		    	   current.p2 = matches[j].primer;
@@ -456,7 +455,6 @@ static void buildPrimerPairsForOneSeq(uint32_t seqid,
 		}
 		 }
 	}
-
    pairs->count=paircount;
 
 }

src/libecoprimer/strictprimers.c

@@ -108,10 +108,11 @@ void addSeqToWordCountTable(pwordcount_t table, uint32_t wordsize, uint32_t circ
 
 	table->inseqcount++;
 
-
-	table->strictcount = ECOREALLOC(table->strictcount,buffersize*sizeof(uint32_t),
+	//fprintf (stderr, "\nOldAddress: %x", table->strictcount);
+	table->strictcount = ECOREALLOC(table->strictcount,(buffersize+5000)*sizeof(uint32_t),
 								"Cannot allocate memory to extend example word count table");
-
+	//fprintf (stderr, " NewAddress: %x\n", table->strictcount);
+	
 	for (i=table->size; i < buffersize; i++)
 		table->strictcount[i]=1;
 
@@ -172,7 +173,7 @@ pwordcount_t lookforStrictPrimer(pecodnadb_t database, uint32_t seqdbsize,
 
     for (i=0;i<seqdbsize;i++)
     {
-    	if (database[i]->isexample)
+    	if (database[i]->isexample && database[i]->SQ_length > options->primer_length)
     	{
 
     		if (first)

src/libecoprimer/taxstats.c

@@ -6,10 +6,46 @@
  */
 
 #include <search.h>
+//void tdestroy (void *root, void (*free_node)(void *nodep));
+
 #include  "ecoprimer.h"
 
 static int cmptaxon(const void *t1, const void* t2);
 
+void **tree_root = NULL;
+int delete_passes = 0; 
+
+void delete_twalkaction (const void *node, VISIT order, int level)
+{
+	switch (order)
+	{
+	case preorder:
+		delete_passes++;
+		break;
+	case postorder:
+		delete_passes++;
+		break;
+	case endorder:
+		delete_passes++;
+		break;
+	case leaf:
+		if (tree_root)
+			tdelete (node, tree_root,cmptaxon);
+		delete_passes++;
+		break;
+	}
+}
+
+void free_tree_nodes (void *tree)
+{
+	while (1)
+	{
+		delete_passes = 0;
+		twalk (tree, delete_twalkaction);
+		if (delete_passes <= 1) break;
+	}
+}
+
 static int cmptaxon(const void *t1, const void* t2)
 {
 	const size_t taxid1=(size_t)t1;
@@ -35,7 +71,12 @@ int32_t counttaxon(int32_t taxid)
 	if (taxid==-1)
 	{
 		if (taxontree)
+		{
+			tree_root = (void **)&taxontree;
+			//free_tree_nodes (taxontree);
 			ECOFREE(taxontree,"Free taxon tree");
+			tree_root = NULL;
+		}
 		taxontree=NULL;
 		taxoncount=0;
 		return 0;
@@ -97,22 +138,30 @@ int32_t getrankdbstats(pecodnadb_t seqdb, uint32_t seqdbsize, ecotaxonomy_t *tax
 }
 
 
-float taxonomycoverage(ppair_t pair, poptions_t options)
+float taxonomycoverage(ppair_t pair, poptions_t options, pecodnadb_t seqdb,uint32_t seqdbsize)
 {
 	int32_t seqcount;
 	int32_t i;
 	int32_t incount=0;
 	int32_t outcount=0;
+	uint32_t j;
 
 
+	memset (pair->coveredSeqs, 0, seqdbsize*sizeof (int));
 	seqcount=pair->pcr.ampcount;
 
 	counttaxon(-1);
 	for (i=0; i < seqcount; i++)
 		if (pair->pcr.amplifias[i].sequence->isexample
 				&& pair->pcr.amplifias[i].sequence->ranktaxonid > 0 )
+		{
 			incount = counttaxon(pair->pcr.amplifias[i].sequence->ranktaxonid);
 
+			for (j=0; j<seqdbsize; j++)
+				if (pair->pcr.amplifias[i].sequence == seqdb[j])
+					{pair->coveredSeqs[j] = 1; break;}
+		}
+
 	counttaxon(-1);
 	for (i=0; i < seqcount; i++)
 		if (!pair->pcr.amplifias[i].sequence->isexample
@@ -145,12 +194,14 @@ static int cmpamp(const void *ampf1, const void* ampf2)
 	{
 		incr = -1;
 		j = pampf1->length - 1;
+		
 		if (pampf2->strand)
 		{
 			pampf1 = (pamptotaxon_t) ampf2;
 			pampf2 = (pamptotaxon_t) ampf1;
 			chd = 1;
 		}
+		//j = pampf2->length - 1; should have been here and pampf2 instead of pampf1?
 	}
 
 	len = (pampf1->length <= pampf2->length)? pampf1->length: pampf2->length;
@@ -173,6 +224,7 @@ static int cmpamp(const void *ampf1, const void* ampf2)
 	return 0;
 }*/
 
+
 static int cmpamp(const void *ampf1, const void* ampf2)
 {
 	int i;
@@ -183,10 +235,10 @@ static int cmpamp(const void *ampf1, const void* ampf2)
 	char *ch2;
 	int incr1;
 	int incr2;
-	
+
 	pamptotaxon_t pampf1 = (pamptotaxon_t) ampf1;
 	pamptotaxon_t pampf2 = (pamptotaxon_t) ampf2;
-	
+
 	ch1 = pampf1->amplifia;
 	ch2 = pampf2->amplifia;
 	
@@ -218,7 +270,7 @@ static int cmpamp(const void *ampf1, const void* ampf2)
 	
 	if (pampf1->length > pampf2->length) return 1;
 	if (pampf2->length > pampf1->length) return -1;
-	
+
 	return 0;
 }
 
@@ -242,6 +294,8 @@ void taxonomyspecificity (ppair_t pair, pecodnadb_t seqdb,uint32_t seqdbsize)
 	uint32_t i, j;
 	uint32_t ampfindex = 0;
 	int32_t taxid;
+	uint32_t wellidentifiedcount;
+	
 	void *ampftree = NULL;
 	pamptotaxon_t pcurrentampf;
 	pamptotaxon_t *ptmp;
@@ -278,11 +332,14 @@ void taxonomyspecificity (ppair_t pair, pecodnadb_t seqdb,uint32_t seqdbsize)
 	}
 
 	memset (pair->wellIdentifiedSeqs, 0, seqdbsize*sizeof (int));
-	counttaxon(-1);
+	//counttaxon(-1);
 	for (i = 0; i < ampfindex; i++)
 	{
 		if (ampfwithtaxtree[i].taxoncount > 1)
-			twalk(ampfwithtaxtree[i].taxontree, twalkaction);
+		{
+			//printf ("\nampfwithtaxtree[i].taxoncount: %d\n", ampfwithtaxtree[i].taxoncount);
+			//twalk(ampfwithtaxtree[i].taxontree, twalkaction);
+		}
 		//TR 5/9/10 - added code for well identified seqs
 		else if(ampfwithtaxtree[i].taxoncount == 1) /*well identified*/
 		{
@@ -293,6 +350,7 @@ void taxonomyspecificity (ppair_t pair, pecodnadb_t seqdb,uint32_t seqdbsize)
 			{
 				for (j = 0; j < seqdbsize; j++)
 					if (seqdb[j]->ranktaxonid == gtxid 
+							&& seqdb[j]->isexample
 							&&(pair->p1->directCount[j] > 0 
 							|| pair->p1->reverseCount[j] > 0)
 							&& (pair->p2->directCount[j] > 0 
@@ -303,10 +361,18 @@ void taxonomyspecificity (ppair_t pair, pecodnadb_t seqdb,uint32_t seqdbsize)
 			}
 		}
 	}
-
-	pair->notwellidentifiedtaxa = counttaxon(-2);
-	pair->bs = ((float)pair->intaxa - (float)pair->notwellidentifiedtaxa) / pair->intaxa;
-
+	//printf ("\n");
+	counttaxon(-1);
+	wellidentifiedcount = 0;
+	for (j = 0; j < seqdbsize; j++)
+		if (pair->wellIdentifiedSeqs[j] == 1)
+			counttaxon(seqdb[j]->ranktaxonid);
+	wellidentifiedcount = counttaxon(-2);
+	//pair->notwellidentifiedtaxa = counttaxon(-2);
+	pair->notwellidentifiedtaxa = (pair->intaxa-wellidentifiedcount); //counttaxon(-2);
+	//pair->bs = ((float)pair->intaxa - (float)pair->notwellidentifiedtaxa) / pair->intaxa;
+	pair->bs = ((float)wellidentifiedcount) / (float)pair->intaxa;
+	
 	ECOFREE (ampfwithtaxtree, "Free amplifia table");
 
 }