// Copyright 2012 Dmitri Pervouchine (dp@crg.eu), Lab Roderic Guigo

// Bioinformatics and Genomics Group @ Centre for Genomic Regulation

// Parc de Recerca Biomedica: Dr. Aiguader, 88, 08003 Barcelona

//

// This file is a part of the 'bam2ssj' package.

// 'bam2ssj' package is free software: you can redistribute it and/or modify

// it under the terms of the GNU General Public License as published by

// the Free Software Foundation, either version 3 of the License, or

// (at your option) any later version.

//

// 'bam2ssj' package is distributed in the hope that it will be useful,

// but WITHOUT ANY WARRANTY; without even the implied warranty of

// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the

// GNU General Public License for more details.

//

// You should have received a copy of the GNU General Public License

// along with 'bam2ssj' package. If not, see <http://www.gnu.org/licenses/>.

#include <stdio.h>

#include <stdlib.h>

#include <string.h>

#include <time.h>

#include <sys/ioctl.h>

#include <bam.h>

#include "progressbar.h"

#include "list.h"

#define MAXFILEBUFFLENGTH 1000

#define ARRAY_MARGIN 2

#define INFTY 65535

#define WINDOW 1

#define MIN2(A,B) ((A<B) ? (A) : (B))

#define RT_OVRLAP 0

#define RT_GENOME 1

#define RT_KJUNCT 2

#define RT_UJUNCT 3

#define RT_OTHER 4

#define MIN2MAX(A,B) ((A)<(B) ? (double)(A)/(B) : (double)(B)/(A))

#define BAM_UNIQUE_MAP 0x800

#define N_READ_TYPES 5

char read_type_descr[N_READ_TYPES][MAXFILEBUFFLENGTH] = { "genomic overlapping sj", "genomic non-overlapping sj (ignored)",

"known junctions", "unknown junctions (ignored)", "ignored by other reasons"};

typedef struct {

int pos;

int label;

int count00[2];

int count5X[2];

int countX3[2];

list_element* junctions;

} splice_site;

int verbose=1;

int main(int argc,char* argv[]) {

time_t timestamp, current;

int i,j,k;

int a,n;

char *pc;

FILE *input_file;

FILE *output_file;

FILE* log_file=stderr;

bamFile bam_input;

bam_header_t *header;

bam1_t* b;

bam1_core_t *c;

char cps_file_name[MAXFILEBUFFLENGTH]="";

char bam_file_name[MAXFILEBUFFLENGTH]="";

char out_file_name[MAXFILEBUFFLENGTH]="";

char log_file_name[MAXFILEBUFFLENGTH]="";

char buff[MAXFILEBUFFLENGTH];

char chr[MAXFILEBUFFLENGTH];

int beg, beg_prev, end, pos, offset;

int ref_id, ref_id_prev, label;

int s, side;

int read_type, mapped_strand;

char ch;

int limit_counts = 0;

int* contig_count[2];

int* contig_index[2];

splice_site** contig_sites[2];

long int n_reads[N_READ_TYPES][2];

long int n_total_reads = 0;

long int n_skipped_reads = 0;

int max_intron_length=0;

int min_intron_length=0;

int ignore_gene_labels = 0;

int stranded = 1;

int rev_compl[2] = {1,0};

int other_end, the_end, donor_id, acceptor_id;

int *cigar;

int flagged = 0;

int margin = 4;

/** reading input from the command line **/

timestamp = time(NULL);

if(argc==1) {

fprintf(stderr, "BAM2SSJ is the utility for fast counting reads covering splice junctions\nCommand line use:\n");

fprintf(stderr, "%s -cps <cps_file> -bam <bam_file> [-out <out_file>] [-log <log_file>] [-maxlen <max_intron_length>] [-minlen <min_intron_length>] [-margin <length>] ",argv[0]);

fprintf(stderr, "[-v suppress verbose output] [-read1 0/1] [-read2 0/1] [-g ignore gene labels] [-u unstranded] [-f count reads flagged 0x800 only]\ntype %s -h for more info\n",argv[0]);

exit(1);

}

for(i=1;i<argc;i++) {

pc = argv[i];

if(*pc == '-') {

if(strcmp(pc+1,"cps") == 0) sscanf(argv[++i], "%s", &cps_file_name[0]);

if(strcmp(pc+1,"bam") == 0) sscanf(argv[++i], "%s", &bam_file_name[0]);

if(strcmp(pc+1,"out") == 0) sscanf(argv[++i], "%s", &out_file_name[0]);

if(strcmp(pc+1,"log") == 0) sscanf(argv[++i], "%s", &log_file_name[0]);

if(strcmp(pc+1,"read1") == 0) sscanf(argv[++i], "%i", &rev_compl[0]);

if(strcmp(pc+1,"read2") == 0) sscanf(argv[++i], "%i", &rev_compl[1]);

if(strcmp(pc+1,"lim") == 0) sscanf(argv[++i], "%i", &limit_counts);

if(strcmp(pc+1,"minlen") == 0) sscanf(argv[++i], "%i", &min_intron_length);

if(strcmp(pc+1,"maxlen") == 0) sscanf(argv[++i], "%i", &max_intron_length);

if(strcmp(pc+1,"margin") == 0) sscanf(argv[++i], "%i", &margin);

if(strcmp(pc+1,"v") == 0) verbose = 0;

if(strcmp(pc+1,"g") == 0) ignore_gene_labels = 1;

if(strcmp(pc+1,"u") == 0) stranded = 0;

if(strcmp(pc+1,"f") == 0) flagged = 1;

if(strcmp(pc+1,"h") ==0 ) {

fprintf(stderr, "Input: (1) sorted BAM file\n");

fprintf(stderr, "\t(2) CPS (chromosome-position-strand) tab-delimited file sorted by position (chr1 100 + etc)\n\n");

fprintf(stderr, "\tIn order to get CPS file from gtf, use the utility gtf2cps.sh\n");

fprintf(stderr, "\tImportant: CPS must be sorted by position ONLY!\n\n");

fprintf(stderr, "\tIf the 4th column contains (a numeric) gene label then only splice junctions within the same gene will be considered (unless the '-g' option is active)\n");

fprintf(stderr, "\tThe utility to generate CPS with gene labels is gtf2cps_with_gene_id.sh (or update the script accordingly if you are using genome other than human)\n\n");

fprintf(stderr, "Options:\n");

fprintf(stderr, "\t-maxlen <upper limit on intron length>; 0 = no limit (default=%i)\n",max_intron_length);

fprintf(stderr, "\t-minlen <lower limit on intron length>; 0 = no limit (default=%i)\n",min_intron_length);

fprintf(stderr, "\t-margin <length> minimum number of flanking nucleotides in the read in order to support SJ or cover EB, (default=%i)\n",margin);

fprintf(stderr, "\t-read1 0/1, reverse complement read1 no/yes (default=%i)\n",rev_compl[0]);

fprintf(stderr, "\t-read2 0/1, reverse complement read2 no/yes (default=%i)\n",rev_compl[1]);

fprintf(stderr, "\t-g ignore gene labels (column 4 of cps), default=%s\n", ignore_gene_labels ? "ON" : "OFF");

fprintf(stderr, "\t-u ignore strand (all reads map to the correct strand), default=%s\n", stranded ? "OFF" : "ON");

fprintf(stderr, "\t-f count only reads that are flagged 0x800 (uniquely mapped reads), default=%s\n", flagged ? "ON" : "OFF");

fprintf(stderr, "Output: tab-delimited (default=stdout)\n");

fprintf(stderr, "\tColumn 1 is splice_junction_id\n");

fprintf(stderr, "\tColumns 2-6 are counts of 53, 5X, X3, 50, and 03 reads for the correct (annotated) strand\n");

fprintf(stderr, "\tColumns 7-11 are similar counts for the incorrect (opposite to annotated) strand\n");

fprintf(stderr, "Descriptive read statistics are reported to stderr\n");

exit(1);

}

}

}

if(log_file_name[0]==0) {

log_file = stderr;

}

else {

log_file = fopen(log_file_name,"w");

if(log_file == NULL) log_file = stderr;

}

if(bam_file_name[0]==0) {

fprintf(log_file,"Bam not specified, exiting\n");

exit(1);

}

if(cps_file_name[0]==0) {

fprintf(log_file,"Input not specified, exiting\n");

exit(1);

}

if(out_file_name[0]==0) {

fprintf(log_file,"[Warning: output set to stdout]\n");

output_file = stdout;

}

else {

output_file = fopen(out_file_name,"w");

if(output_file == NULL) {

fprintf(log_file,"[Warning: output set to stdout]\n");

output_file = stdout;

}

}

if(max_intron_length>0) {

if(verbose) fprintf(log_file,"[Warning: set max intron length=%i]\n",max_intron_length);

}

if(ignore_gene_labels) {

if(verbose) fprintf(log_file,"[Warning: ignoring gene labels (column 4)]\n");

}

if(flagged) {

if(verbose) fprintf(log_file,"[Warning: only look at reads flagged 0x800]\n");

}

if(margin>0) {

if(verbose) fprintf(log_file,"[Warning: read margin set to %i]\n", margin);

}

if(verbose) {

for(s = 0; s < 2; s++) if(rev_compl[s]) fprintf(log_file,"[Warning: take reverse complement of read %i]\n", s+1);

fprintf(log_file,"[Warning: stranded = %s]\n", stranded ? "TRUE" : "FALSE (always correct strand)");

if(ignore_gene_labels) fprintf(log_file,"[Warning: ignore gene labels (column 4)]\n");

}

for(i = 0; i < N_READ_TYPES; i++) for(s = 0; s < 2; s++) n_reads[i][s] = 0;

/** initatializing BAM and header **/

bam_input = bam_open(bam_file_name, "r");

header = bam_header_read(bam_input);

if(bam_input == NULL || header == NULL) {

fprintf(log_file,"BAM can't be opened or contains no header, exiting\n");

exit(1);

}

/** reading input from CPS **/

input_file = fopen(cps_file_name, "r");

if(input_file == NULL) {

fprintf(log_file,"CPS can't be opened, exiting\n");

exit(1);

}

/** populating gene structure arrays **/

for(s = 0; s < 2; s++) {

contig_count[s] = (int*) malloc(sizeof(int) * (header->n_targets + ARRAY_MARGIN));

contig_index[s] = (int*) malloc(sizeof(int) * (header->n_targets + ARRAY_MARGIN));

contig_sites[s] = (splice_site**) malloc(sizeof(splice_site*) * (header->n_targets + ARRAY_MARGIN));

if(contig_count[s] == NULL || contig_sites[s] == NULL || contig_index[s] == NULL) {

fprintf(log_file, "Not enought memory, exiting\n");

exit(1);

}

}

for(s = 0; s < 2; s++)

for(i=0; i < header->n_targets; i++)

contig_count[s][i] = contig_index[s][i] = 0;

if(verbose) fprintf(log_file, "Reading %s pass1", cps_file_name);

while(fgets(buff, MAXFILEBUFFLENGTH, input_file)) {

sscanf(buff, "%s %*i %c", &chr[0], &ch);

bam_parse_region(header, chr, &i, &beg, &end);

s = (ch == '+' ? 0 : 1);

if(i < header->n_targets && i>=0) contig_count[s][i]++;

}

for(s = 0; s < 2; s++) {

for(i = 0;i < header->n_targets; i++) {

contig_sites[s][i] = (splice_site*) malloc(sizeof(splice_site) * (contig_count[s][i] + ARRAY_MARGIN));

if(contig_sites[s][i] == NULL) {

fprintf(log_file, "Not enought memory, exiting\n");

exit(1);

}

}

}

if(verbose) fprintf(log_file, "\n");

if(verbose) fprintf(log_file, "Reading %s pass2",cps_file_name);

fseek(input_file, 0, SEEK_SET);

while(fgets(buff, MAXFILEBUFFLENGTH, input_file)) {

sscanf(buff, "%s %i %c %i", &chr[0], &pos, &ch, &label);

bam_parse_region(header, chr, &i, &beg, &end);

s = (ch == '+' ? 0 : 1);

if(i < header->n_targets && i>=0) {

if(contig_index[s][i]>0) {

if(pos < contig_sites[s][i][contig_index[s][i]-1].pos) {

fprintf(log_file, "Splice sites weren't sorted, exiting\n");

exit(1);

}

}

contig_sites[s][i][contig_index[s][i]].pos = pos;

contig_sites[s][i][contig_index[s][i]].label = ignore_gene_labels ? 0 : label;

for(side = 0; side < 2; side++) {

contig_sites[s][i][contig_index[s][i]].count00[side] = 0;

contig_sites[s][i][contig_index[s][i]].count5X[side] = 0;

contig_sites[s][i][contig_index[s][i]].countX3[side] = 0;

contig_sites[s][i][contig_index[s][i]].junctions = NULL;

}

contig_index[s][i]++;

}

}

if(verbose) fprintf(log_file, "\n");

for(s = 0; s < 2; s++)

for(i = 0;i < header->n_targets; i++)

contig_index[s][i] = 0;

/** analysis starts here **/

b = bam_init1();

k = 0;

ref_id_prev = -1;

beg_prev = -1;

while(bam_read1(bam_input, b)>=0) {

c = &b->core;

ref_id = c->tid;

if(ref_id<0) continue;

if(flagged && ((c->flag & 0x800) == 0)) {

n_skipped_reads++;

continue;

}

if(stranded && ((c->flag & BAM_FREAD1) && (c->flag & BAM_FREAD2) || !(c->flag & BAM_FREAD1) && !(c->flag & BAM_FREAD2))) {

n_skipped_reads++;

continue;

}

cigar = bam1_cigar(b);

if(ref_id != ref_id_prev && ref_id_prev >= 0) {

if(contig_index[0][ref_id_prev] + contig_index[1][ref_id_prev] < contig_count[0][ref_id_prev] + contig_count[1][ref_id_prev]) {

if(log_file==stderr) progressbar(1, 1, header->target_name[ref_id_prev], verbose);

}

beg_prev = -1;

}

/*if(ref_id < ref_id_prev) {

fprintf(log_file,"BAM file wasn't sorted, exiting\n");

exit(1);

}*/

ref_id_prev = ref_id;

beg = c->pos + 1;

if(beg < beg_prev) {

fprintf(log_file,"BAM file wasn't sorted, exiting\n");

exit(1);

}

beg_prev = beg;

s = ((c->flag & BAM_FREVERSE)>0);

mapped_strand = (c->flag & BAM_FREAD1) ? (s + rev_compl[0]) & 1 : (s + rev_compl[1]) & 1;

the_end = bam_calend(c, cigar);

for(s = 0; s < 1 + stranded; s++) {

end = beg;

side = (s == mapped_strand) ? 0 : 1;

side *= stranded;

// keep reading until the currect site is on the same chromosome downstream of the read

while(contig_sites[s][ref_id][contig_index[s][ref_id]].pos < beg && contig_index[s][ref_id] < contig_count[s][ref_id]) {

contig_index[s][ref_id]++;

if(log_file==stderr) progressbar(contig_index[0][ref_id]+contig_index[1][ref_id], contig_count[0][ref_id]+contig_count[1][ref_id], header->target_name[ref_id], verbose);

}

read_type = RT_OTHER;

if(contig_index[s][ref_id]<contig_count[s][ref_id]) {

// check if the read is a split read and find its other end

read_type = RT_GENOME;

for(i = 0; i < c->n_cigar; i++) {

offset = cigar[i] >> 4;

switch(cigar[i] & 0x0F) {

case BAM_CMATCH: end += offset; // match to the reference

break;

case BAM_CINS: end += 0; // insertion to the reference, pointer stays unchanged

break;

case BAM_CDEL: end += offset; // deletion from the reference (technically the same as 'N') pointer moves

break;

case BAM_CREF_SKIP: other_end = end + offset;

donor_id = acceptor_id = -INFTY;

if(end - beg < margin) break;

if(the_end - other_end < margin) break;

for(j = contig_index[s][ref_id]; contig_sites[s][ref_id][j].pos <= other_end && j < contig_count[s][ref_id];j++) {

if(contig_sites[s][ref_id][j].pos - end < min_intron_length && min_intron_length > 0) continue;

if(contig_sites[s][ref_id][j].pos - end > max_intron_length && max_intron_length > 0) break;

if(contig_sites[s][ref_id][j].label == contig_sites[s][ref_id][contig_index[s][ref_id]].label) {

if(contig_sites[s][ref_id][j].pos == end - 1) donor_id = j;

if(contig_sites[s][ref_id][j].pos == other_end) acceptor_id = j;

}

}

if(donor_id>0 && acceptor_id>0) {

update_count(&contig_sites[s][ref_id][donor_id].junctions, acceptor_id, side);

contig_sites[s][ref_id][donor_id].count5X[side]++;

contig_sites[s][ref_id][acceptor_id].countX3[side]++;

read_type = RT_KJUNCT;

}

else {

read_type = RT_UJUNCT;

}

end = other_end;

break;

case BAM_CSOFT_CLIP:

case BAM_CHARD_CLIP:

case BAM_CPAD: break;

default: read_type = RT_OTHER;

}

}

if(read_type == RT_GENOME) {

for(j=contig_index[s][ref_id]; beg + margin <= contig_sites[s][ref_id][j].pos && contig_sites[s][ref_id][j].pos < end - margin && j<contig_count[s][ref_id]; j++) {

contig_sites[s][ref_id][j].count00[side]++;

read_type = RT_OVRLAP;

k++;

}

}

}

n_reads[read_type][side]++;

}

n_total_reads++;

if(k>limit_counts && limit_counts>0) break;

}

if(ref_id_prev > 0) {

if(contig_index[0][ref_id_prev] + contig_index[1][ref_id_prev] < contig_count[0][ref_id_prev] + contig_count[1][ref_id_prev]) {

if(log_file==stderr) progressbar(1, 1, header->target_name[ref_id_prev], verbose);

}

}

/**** output here *****/

for(s = 0; s < 2; s++) {

for(i = 0; i < header->n_targets; i++) {

for(j = 0; j < contig_count[s][i]; j++) {

list_element* ptr = contig_sites[s][i][j].junctions;

while(ptr!=NULL) {

fprintf(output_file, "%s_%i_%i_%i", header->target_name[i], contig_sites[s][i][j].pos + 1, contig_sites[s][i][ptr->id].pos - 1,(s == 0) ? 1 : -1);

for(side = 0; side < 1 + stranded; side++) {

fprintf(output_file, "\t%i\t%i\t%i\t%i\t%i", ptr->count53[side],

contig_sites[s][i][j].count5X[side] - ptr->count53[side],

contig_sites[s][i][ptr->id].countX3[side] - ptr->count53[side],

contig_sites[s][i][j].count00[side],

contig_sites[s][i][ptr->id].count00[side]);

}

fprintf(output_file, "\n");

ptr= ptr->next;

}

}

}

}

for(s = 0; s < 2; s++) {

for(i = 0;i < header->n_targets; i++) {

for(j = 0; j < contig_count[s][i]; j++) {

destroy_list(contig_sites[s][i][j].junctions);

}

free(contig_sites[s][i]);

}

free(contig_sites[s]);

free(contig_index[s]);

free(contig_count[s]);

}

bam_header_destroy(header);

bam_close(bam_input);

bam_destroy1(b);

fclose(input_file);

fclose(output_file);

if(verbose) fprintf(log_file,"\n");

current = time(NULL);

fprintf(log_file,"Read statistics for %s\n",bam_file_name);

for(s = 0; s < 1 + stranded; s++) fprintf(log_file,"%16s", s == 0 ? "correct" : "incorrect");

if(stranded) fprintf(log_file,"\tmin/max");

fprintf(log_file,"\tstrand\n");

for(i = 0; i < N_READ_TYPES; i++) {

for(s = 0; s < 1 + stranded; s++) fprintf(log_file,"%16li", n_reads[i][s]);

if(stranded) fprintf(log_file,"\t%1.2lf",MIN2MAX(n_reads[i][1],n_reads[i][0]));

fprintf(log_file,"\t%s\n",read_type_descr[i]);

}

for(s = 0; s < 1 + stranded; s++) fprintf(log_file,"%16li", n_skipped_reads);

if(stranded) fprintf(log_file,"\t");

fprintf(log_file,"\tskipped\n");

for(s = 0; s < 1 + stranded; s++) fprintf(log_file,"%16li", n_total_reads + n_skipped_reads);

if(stranded) fprintf(log_file,"\t");

fprintf(log_file,"\ttotal reads\n");

fprintf(log_file,"Completed in %1.0lf seconds\n",difftime(current,timestamp));

return 0;

}