/* No gap is expected in the longer sequence */
void nw_test_no_gaps_in_longer()
{
nw_aligner_t *nw = needleman_wunsch_new();
alignment_t *aln = alignment_create(256);
const char* seq_a = "aaaaacg";
const char* seq_b = "acgt";
int match = 1;
int mismatch = -2;
int gap_open = -4;
int gap_extend = -1;
bool no_start_gap_penalty = false, no_end_gap_penalty = false;
bool no_gaps_in_a = true, no_gaps_in_b = false;
bool no_mismatches = false, case_sensitive = true;
scoring_t scoring;
scoring_init(&scoring, match, mismatch, gap_open, gap_extend,
no_start_gap_penalty, no_end_gap_penalty,
no_gaps_in_a, no_gaps_in_b,
no_mismatches, case_sensitive);
needleman_wunsch_align(seq_a, seq_b, &scoring, nw, aln);
// ASSERT(strcmp(aln->result_a, "aaaaacg") == 0 &&
// strcmp(aln->result_b, "acgt---") == 0);
ASSERT(strcmp(aln->result_a, "aaaaacg-") == 0 &&
strcmp(aln->result_b, "a----cgt") == 0);
alignment_free(aln);
needleman_wunsch_free(nw);
}
/* First sequence is aligned to the corresponding (equal) substring of the second
* sequence because both gaps at start and at end are free */
void nw_test_free_gaps_at_ends()
{
nw_aligner_t *nw = needleman_wunsch_new();
alignment_t *result = alignment_create(256);
const char* seq_a = "acg";
const char* seq_b = "tttacgttt";
int match = 1;
int mismatch = -1;
int gap_open = -4;
int gap_extend = -1;
bool no_start_gap_penalty = true, no_end_gap_penalty = true;
bool no_gaps_in_a = false, no_gaps_in_b = false;
bool no_mismatches = false, case_sensitive = true;
scoring_t scoring;
scoring_init(&scoring, match, mismatch, gap_open, gap_extend,
no_start_gap_penalty, no_end_gap_penalty,
no_gaps_in_a, no_gaps_in_b,
no_mismatches, case_sensitive);
needleman_wunsch_align(seq_a, seq_b, &scoring, nw, result);
ASSERT(strcmp(result->result_a, "---acg---") == 0 &&
strcmp(result->result_b, "tttacgttt") == 0);
alignment_free(result);
needleman_wunsch_free(nw);
}
void nw_test_no_mismatches()
{
nw_aligner_t *nw = needleman_wunsch_new();
alignment_t *result = alignment_create(256);
int match = 1;
int mismatch = -2;
int gap_open = -4;
int gap_extend = -1;
bool no_start_gap_penalty = false, no_end_gap_penalty = false;
bool no_gaps_in_a = false, no_gaps_in_b = false;
bool no_mismatches = true, case_sensitive = true;
scoring_t scoring;
scoring_init(&scoring, match, mismatch, gap_open, gap_extend,
no_start_gap_penalty, no_end_gap_penalty,
no_gaps_in_a, no_gaps_in_b,
no_mismatches, case_sensitive);
needleman_wunsch_align("atc", "ac", &scoring, nw, result);
ASSERT(strcmp(result->result_a, "atc") == 0 &&
strcmp(result->result_b, "a-c") == 0);
needleman_wunsch_align("cgatcga", "catcctcga", &scoring, nw, result);
ASSERT(strcmp(result->result_a, "cgatc---ga") == 0 &&
strcmp(result->result_b, "c-atcctcga") == 0);
alignment_free(result);
needleman_wunsch_free(nw);
}
// Setup pairwise aligner
static void nw_aligner_setup()
{
nw_aligner = needleman_wunsch_new();
aln = alignment_create(1024);
scoring_init(&nw_scoring_flank, nwmatch, nwmismatch, nwgapopen, nwgapextend,
true, true, 0, 0, 0, 0);
scoring_init(&nw_scoring_allele, nwmatch, nwmismatch, nwgapopen, nwgapextend,
false, false, 0, 0, 0, 0);
}
void align(char* seq_a, char* seq_b)
{
// Variables to store alignment result
sw_aligner_t *sw = smith_waterman_new();
alignment_t *result = alignment_create(256);
// Decide on scoring
int match = 1;
int mismatch = -2;
int gap_open = -4;
int gap_extend = -1;
// Don't penalise gaps at the start
// ACGATTT
// ----TTT would score +3 (when match=+1)
char no_start_gap_penalty = 1;
// ..or gaps at the end e.g.
// ACGATTT
// ACGA--- would score +4 (when match=+1)
char no_end_gap_penalty = 1;
char no_gaps_in_a = 0, no_gaps_in_b = 0;
char no_mismatches = 0;
// Compare character case-sensitively (usually set to 0 for DNA etc)
char case_sensitive = 0;
scoring_t scoring;
scoring_init(&scoring, match, mismatch, gap_open, gap_extend,
no_start_gap_penalty, no_end_gap_penalty,
no_gaps_in_a, no_gaps_in_b, no_mismatches, case_sensitive);
// Add some special cases
// x -> y means x in seq1 changing to y in seq2
scoring_add_mutation(&scoring, 'a', 'c', -2); // a -> c give substitution score -2
scoring_add_mutation(&scoring, 'c', 'a', -1); // c -> a give substitution score -1
// We could also prohibit the aligning of characters not given as special cases
// scoring.use_match_mismatch = 0;
smith_waterman_align(seq_a, seq_b, &scoring, sw);
while(smith_waterman_fetch(sw, result))
{
printf("seqA: %s [start:%zu]\n", result->result_a, result->pos_a);
printf("seqB: %s [start:%zu]\n", result->result_b, result->pos_b);
printf("alignment score: %i\n\n", result->score);
}
// Free memory for storing alignment results
smith_waterman_free(sw);
alignment_free(result);
}
CallDecomp* call_decomp_init(htsFile *vcffh, bcf_hdr_t *vcfhdr)
{
CallDecomp *dc = ctx_calloc(1, sizeof(CallDecomp));
dc->nw_aligner = needleman_wunsch_new();
dc->aln = alignment_create(1024);
dc->scoring = ctx_calloc(1, sizeof(dc->scoring[0]));
scoring_system_default(dc->scoring);
dc->vcffh = vcffh;
dc->vcfhdr = vcfhdr;
dc->v = bcf_init();
strbuf_alloc(&dc->sbuf, 256);
return dc;
}
void nw_test_no_mismatches_rand()
{
nw_aligner_t *nw = needleman_wunsch_new();
alignment_t *aln = alignment_create(256);
int match = 1;
int mismatch = -2;
int gap_open = -4;
int gap_extend = -1;
bool no_start_gap_penalty = false, no_end_gap_penalty = false;
bool no_gaps_in_a = false, no_gaps_in_b = false;
bool no_mismatches = true, case_sensitive = true;
scoring_t scoring;
scoring_init(&scoring, match, mismatch, gap_open, gap_extend,
no_start_gap_penalty, no_end_gap_penalty,
no_gaps_in_a, no_gaps_in_b,
no_mismatches, case_sensitive);
char seqa[100], seqb[100];
size_t i;
// Run 50 random alignments
for(i = 0; i < 50; i++)
{
make_rand_seq(seqa, sizeof(seqa));
make_rand_seq(seqb, sizeof(seqb));
needleman_wunsch_align(seqa, seqb, &scoring, nw, aln);
// Check no mismatches
char *a = aln->result_a, *b = aln->result_b;
while(1) {
ASSERT(*a == '-' || *b == '-' || *a == *b);
// printf("Seq: '%s' '%s'\n", aln->result_a, aln->result_b);
// exit(EXIT_FAILURE);
if(!*a && !*b) break;
a++; b++;
}
}
alignment_free(aln);
needleman_wunsch_free(nw);
}
int main(int argc, char* argv[])
{
#ifdef SEQ_ALIGN_VERBOSE
printf("VERBOSE: on\n");
#endif
sw_set_default_scoring();
cmd = cmdline_new(argc, argv, &scoring, SEQ_ALIGN_SW_CMD);
// Align!
sw = smith_waterman_new();
result = alignment_create(256);
if(cmd->seq1 != NULL)
{
// Align seq1 and seq2
align(cmd->seq1, cmd->seq2, NULL, NULL);
}
// Align from files
size_t i, num_of_file_pairs = cmdline_get_num_of_file_pairs(cmd);
for(i = 0; i < num_of_file_pairs; i++)
{
const char *file1 = cmdline_get_file1(cmd, i);
const char *file2 = cmdline_get_file2(cmd, i);
if(file1 != NULL && *file1 == '\0' && file2 == NULL) {
wait_on_keystroke = 1;
file1 = "-";
}
align_from_file(file1, file2, &align_pair_from_file, !cmd->interactive);
}
// Free memory for storing alignment results
smith_waterman_free(sw);
alignment_free(result);
cmdline_free(cmd);
return EXIT_SUCCESS;
}
int main(int argc, char **argv)
{
if(argc != 2) print_usage(argv);
char *seq = argv[1];
size_t seqlen = strlen(seq);
// Go
int match = 1, mismatch = -1, gap_open = -4, gap_extend = -1;
bool no_start_gap_penalty = false, no_end_gap_penalty = false;
bool no_gaps_in_a = true, no_gaps_in_b = true;
bool no_mismatches = true, case_sensitive = true;
scoring_t scoring;
scoring_init(&scoring, match, mismatch, gap_open, gap_extend,
no_start_gap_penalty, no_end_gap_penalty,
no_gaps_in_a, no_gaps_in_b,
no_mismatches, case_sensitive);
// Alignment results stored here
sw_aligner_t *sw = smith_waterman_new();
alignment_t *aln = alignment_create(seqlen+1);
smith_waterman_align(seq, seq, &scoring, sw);
// Loop over results
while(smith_waterman_fetch(sw, aln))
{
if(aln->pos_a < aln->pos_b) {
fputs(aln->result_a, stdout);
printf(" [%zu,%zu]\n", aln->pos_a, aln->pos_b);
}
}
smith_waterman_free(sw);
alignment_free(aln);
return EXIT_SUCCESS;
}
void sw_test_no_gaps_smith_waterman()
{
sw_aligner_t *sw = smith_waterman_new();
alignment_t *result = alignment_create(256);
const char* seq_a = "gacag";
const char* seq_b = "tgaagt";
int match = 1;
int mismatch = -2;
int gap_open = -4;
int gap_extend = -1;
bool no_start_gap_penalty = false, no_end_gap_penalty = false;
bool no_gaps_in_a = true, no_gaps_in_b = true;
bool no_mismatches = false, case_sensitive = true;
scoring_t scoring;
scoring_init(&scoring, match, mismatch, gap_open, gap_extend,
no_start_gap_penalty, no_end_gap_penalty,
no_gaps_in_a, no_gaps_in_b,
no_mismatches, case_sensitive);
smith_waterman_align(seq_a, seq_b, &scoring, sw);
smith_waterman_fetch(sw, result);
ASSERT(strcmp(result->result_a, "ga") == 0 &&
strcmp(result->result_b, "ga") == 0);
smith_waterman_fetch(sw, result);
ASSERT(strcmp(result->result_a, "ag") == 0 &&
strcmp(result->result_b, "ag") == 0);
alignment_free(result);
smith_waterman_free(sw);
}
static PyObject * nw_align_wrapper(PyObject *self, PyObject *args, PyObject *kw)
{
const char *seq1, *seq2;
// Decide on scoring
int match = 1;
int mismatch = -2;
int gap_open = -4;
int gap_extend = -1;
// Don't penalise gaps at the start
// ACGATTT
// ----TTT would score +3 (when match=+1)
int no_start_gap_penalty = 0;
// ..or gaps at the end e.g.
// ACGATTT
// ACGA--- would score +4 (when match=+1)
int no_end_gap_penalty = 0;
int no_gaps_in_a = 0, no_gaps_in_b = 0;
int no_mismatches = 0;
// Compare character case-sensitively (usually set to 0 for DNA etc)
int case_sensitive = 0;
PyObject * matrix = NULL;
static char *kwlist[] = {"seq1","seq2", "matrix", "match", "mismatch", "gap_open","gap_extend", "no_start_gap_penalty", "no_end_gap_penalty", "no_gaps_in_a", "no_gaps_in_b", "no_mismatches", "case_sensitive", NULL};
PyObject *res = NULL;
if(!PyArg_ParseTupleAndKeywords(args, kw, "ss|Oiiiiiiiiii", kwlist, &seq1, &seq2, &matrix, &match, &mismatch, &gap_open, &gap_extend,
&no_start_gap_penalty, &no_end_gap_penalty, &no_gaps_in_a, &no_gaps_in_b, &no_mismatches, &case_sensitive))
return NULL;
alignment_t *result = alignment_create(256);
// Variables to store alignment result
nw_aligner_t *nw = needleman_wunsch_new();
scoring_t scoring;
scoring_init(&scoring, match, mismatch, gap_open, gap_extend,
no_start_gap_penalty, no_end_gap_penalty,
no_gaps_in_a, no_gaps_in_b, no_mismatches, case_sensitive);
// Add some special cases
// x -> y means x in seq1 changing to y in seq2
if(matrix != NULL)
{
PyObject * mapping = PyMapping_Items(matrix);
if(mapping == NULL)
goto error;
int n = PySequence_Size(mapping);
PyObject *item;
int value;
PyObject *key;
char * char_a;
char * char_b;
int i;
for(i = 0; i < n; i++)
{
item = PySequence_GetItem(mapping, i);
if(item == NULL || !PyTuple_Check(item))
{
Py_XDECREF(item);
Py_DECREF(mapping);
goto error;
}
if(!PyArg_ParseTuple(item, "Oi", &key, &value))
{
PyErr_SetString(PyExc_RuntimeError, "Values of matrix dict should be integers");
Py_XDECREF(item);
Py_DECREF(mapping);
goto error;
}
if(!PyTuple_Check(key))
{
PyErr_SetString(PyExc_RuntimeError, "Keys of matrix dict should be tuples");
Py_XDECREF(item);
Py_DECREF(mapping);
goto error;
}
if(!PyArg_ParseTuple(key, "ss", &char_a, &char_b))
{
PyErr_SetString(PyExc_RuntimeError, "Keys of matrix dict should be tuples with 2 characters as elements.");
Py_XDECREF(item);
Py_DECREF(mapping);
goto error;
}
if(strlen(char_a) != 1 || strlen(char_b) != 1)
{
PyErr_SetString(PyExc_RuntimeError, "Character length should be 1");
Py_XDECREF(item);
Py_DECREF(mapping);
goto error;
}
scoring_add_mutation(&scoring, case_sensitive ? *char_a : tolower(*char_a), case_sensitive ? *char_a : tolower(*char_b), value); // a -> c give substitution score -2
Py_DECREF(item);
}
}
// We could also prohibit the aligning of characters not given as special cases
// scoring.use_match_mismatch = 0;
needleman_wunsch_align(seq1, seq2, &scoring, nw, result);
res = Py_BuildValue("ssi", result->result_a, result->result_b, result->score);
error:
// Free memory for storing alignment results
needleman_wunsch_free(nw);
alignment_free(result);
return res;
}
