void output_updated_edges ( char * outfile )
{
FILE * fp;
char name[256];
unsigned int i, validCounter = 0;
EDGE * edge;
sprintf ( name, "%s.updated.edge", outfile );
fp = ckopen ( name, "w" );
for ( i = 1; i <= num_ed; i++ )
{ validCounter++; }
fprintf ( fp, "EDGEs %d\n", validCounter );
validCounter = 0;
for ( i = 1; i <= num_ed; i++ )
{
edge = &edge_array[i];
fprintf ( fp, ">length %d,", edge->length );
print_kmer ( fp, vt_array[edge->from_vt].kmer, ',' );
print_kmer ( fp, vt_array[edge->to_vt].kmer, ',' );
if ( EdSmallerThanTwin ( i ) )
{ fprintf ( fp, "1," ); }
else if ( EdLargerThanTwin ( i ) )
{ fprintf ( fp, "-1," ); }
else
{ fprintf ( fp, "0," ); }
fprintf ( fp, "%d\n", edge->cvg );
}
fclose ( fp );
}
void output_contig ( EDGE * ed_array, unsigned int ed_num, char * outfile, int cut_len )
{
char temp[256];
FILE * fp, *fp_contig;
int flag, count, len_c;
int signI;
unsigned int i;
long long sum = 0, N90, N50;
unsigned int * length_array;
boolean tip;
sprintf ( temp, "%s.contig", outfile );
fp = ckopen ( temp, "w" );
qsort ( &ed_array[1], ed_num, sizeof ( EDGE ), cmp_edge );
length_array = ( unsigned int * ) ckalloc ( ed_num * sizeof ( unsigned int ) );
kmerSeq = ( char * ) ckalloc ( overlaplen * sizeof ( char ) );
//first scan for number counting
count = len_c = 0;
for ( i = 1; i <= ed_num; i++ )
{
if ( ( ed_array[i].length + overlaplen ) >= len_bar )
{ length_array[len_c++] = ed_array[i].length + overlaplen; }
if ( ed_array[i].length < 1 || ed_array[i].deleted )
{ continue; }
count++;
if ( EdSmallerThanTwin ( i ) )
{ i++; }
}
sum = 0;
for ( signI = len_c - 1; signI >= 0; signI-- )
{ sum += length_array[signI]; }
if ( len_c > 0 )
{ printf ( "%d ctgs longer than %d, sum up %lldbp, with average length %lld\n", len_c, len_bar, sum, sum / len_c ); }
qsort ( length_array, len_c, sizeof ( length_array[0] ), cmp_int );
printf ( "the longest is %dbp, ", length_array[len_c - 1] );
N50 = sum * 0.5;
N90 = sum * 0.9;
sum = flag = 0;
for ( signI = len_c - 1; signI >= 0; signI-- )
{
sum += length_array[signI];
if ( !flag && sum >= N50 )
{
printf ( "contig N50 is %d bp,", length_array[signI] );
flag = 1;
}
if ( sum >= N90 )
{
printf ( "contig N90 is %d bp\n", length_array[signI] );
break;
}
}
//fprintf(fp,"Number %d\n",count);
for ( i = 1; i <= ed_num; i++ )
{
//if(ed_array[i].multi!=1||ed_array[i].length<1||(ed_array[i].length+overlaplen)<cut_len)
if ( ed_array[i].deleted || ed_array[i].length < 1 )
{ continue; }
if ( ed_array[i].arcs && ed_array[getTwinEdge ( i )].arcs )
{ tip = 0; }
else
{ tip = 1; }
output_1contig ( i, & ( ed_array[i] ), fp, tip );
if ( EdSmallerThanTwin ( i ) )
{ i++; }
}
fclose ( fp );
free ( ( void * ) kmerSeq );
free ( ( void * ) length_array );
printf ( "%d contigs longer than %d output\n", count, cut_len );
sprintf ( temp, "%s.ContigIndex", outfile );
fp_contig = ckopen ( temp, "w" );
fprintf ( fp_contig, "Edge_num %d %d\n", ed_num, count );
fprintf ( fp_contig, "index\tlength\treverseComplement\n" );
for ( i = 1; i <= num_ed; i++ )
{
fprintf ( fp_contig, "%d\t%d\t", i, edge_array[i].length + overlaplen );
if ( EdSmallerThanTwin ( i ) )
{
fprintf ( fp_contig, "1\n" );
i++;
}
else if ( EdLargerThanTwin ( i ) )
{ fprintf ( fp_contig, "-1\n" ); }
else
{ fprintf ( fp_contig, "0\n" ); }
}
fclose ( fp_contig );
}
void output_contig (EDGE * ed_array, unsigned int ed_num, char *outfile, int cut_len)
{
char temp[256];
FILE * fp, *fp_contig;
int flag, count, len_c;
int signI;
unsigned int i, j, diff_len=0;
long long sum = 0, N90, N50;
unsigned int *length_array;
boolean tip;
sprintf (temp, "%s.contig", outfile);
fp = ckopen (temp, "w");
index_array = (unsigned int *)ckalloc((ed_num+1)*sizeof(unsigned int));
unsigned int * all_length_arr = (unsigned int*) ckalloc((ed_num+1)*sizeof(unsigned int));
flag_array = (unsigned int*)ckalloc((ed_num+1)*sizeof(unsigned int));
for (i=1; i<=ed_num; ++i)
{
index_array[i] = ed_array[i].length;
all_length_arr[i] = ed_array[i].length;
}
qsort(&all_length_arr[1], ed_num, sizeof(all_length_arr[0]), cmp_int);
for (i=1; i<=ed_num; ++i)
{
for (j=i+1; j<=ed_num; ++j)
{
if (all_length_arr[i] != all_length_arr[j])
break;
}
all_length_arr[++diff_len] = all_length_arr[i];
flag_array[diff_len] = i;
i = j-1;
}
for (i=1; i<=ed_num; ++i)
{
index_array[i] = uniqueLenSearch(all_length_arr, flag_array, diff_len, index_array[i]);
}
for (i=1; i<=ed_num; ++i)
{
flag_array[index_array[i]] = i;
}
free((void*)all_length_arr);
length_array = (unsigned int *) ckalloc (ed_num * sizeof (unsigned int));
kmerSeq = (char *) ckalloc (overlaplen * sizeof (char));
count = len_c = 0;
for (i = 1; i <= ed_num; i++)
{
if ((ed_array[i].length + overlaplen) >= len_bar)
{
length_array[len_c++] = ed_array[i].length + overlaplen;
}
if (ed_array[i].length < 1 || ed_array[i].deleted)
{
continue;
}
count++;
if (EdSmallerThanTwin (i))
{
i++;
}
}
sum = 0;
for (signI = len_c - 1; signI >= 0; signI--)
{
sum += length_array[signI];
}
qsort ( length_array, len_c, sizeof ( length_array[0] ), cmp_int );
if ( len_c > 0 )
{
printf ( "%d ctgs longer than %d, sum up %lldbp, with average length %lld\n", len_c, len_bar, sum, sum / len_c );
printf ( "the longest is %dbp, ", length_array[len_c - 1] );
}
N50 = sum * 0.5;
N90 = sum * 0.9;
sum = flag = 0;
for (signI = len_c - 1; signI >= 0; signI--)
{
sum += length_array[signI];
if (!flag && sum >= N50)
{
printf ("contig N50 is %d bp,", length_array[signI]);
flag = 1;
}
if (sum >= N90)
{
printf ("contig N90 is %d bp\n", length_array[signI]);
break;
}
}
for (i = 1; i <= ed_num; i++)
{
j = flag_array[i];
if (ed_array[j].deleted || ed_array[j].length < 1)
{
continue;
}
if (ed_array[j].arcs && ed_array[getTwinEdge (j)].arcs)
{
tip = 0;
}
else
{
tip = 1;
}
output_1contig (i, &(ed_array[j]), fp, tip);
if (EdSmallerThanTwin (j))
{
i++;
}
}
fclose (fp);
free ((void *) kmerSeq);
free ((void *) length_array);
printf ("%d contigs longer than %d output\n", count, cut_len);
sprintf (temp, "%s.ContigIndex", outfile);
fp_contig = ckopen (temp, "w");
fprintf (fp_contig, "Edge_num %d %d\n", ed_num, count);
fprintf (fp_contig, "index\tlength\treverseComplement\n");
for (i = 1; i <= num_ed; i++)
{
j = flag_array[i];
fprintf (fp_contig, "%d\t%d\t", i, edge_array[j].length + overlaplen);
if (EdSmallerThanTwin (j))
{
fprintf (fp_contig, "1\n");
i++;
}
else if (EdLargerThanTwin (j))
{
fprintf (fp_contig, "-1\n");
}
else
{
fprintf (fp_contig, "0\n");
}
}
fclose (fp_contig);
}
