xxay/hifiasm
1
0
Fork 0
mirror of https://github.com/chhylp123/hifiasm.git synced 2026-09-25 03:18:11 +08:00
Code Issues Packages Projects Releases Wiki Activity

r201: automatically adjust -N

Browse Source
This commit is contained in:
Heng Li
2020-04-05 22:43:01 -04:00
parent 3773610532
commit adbc65feda
5 changed files with 43 additions and 33 deletions
Download Patch File Download Diff File Expand all files Collapse all files
CommandLines.cpp
+17 -14
View File
@@ -35,15 +35,11 @@ void Print_H(hifiasm_opt_t* asm_opt)
fprintf(stderr, " -w INT minimizer window size [%d]\n", asm_opt->mz_win);
fprintf(stderr, " -f INT number of bits for bloom filter [%d]\n", asm_opt->bf_shift);
fprintf(stderr, " -D FLOAT drop k-mers occuring >FLOAT*coverage times [%.1f]\n", asm_opt->high_factor);
fprintf(stderr, " -N INT consider up to INT overlaps for each oriented read [%d]\n", asm_opt->max_n_chain);
///fprintf(stderr, " -l load all overlaps from disk, can avoid overlap calculation [%d]\n", asm_opt->load_index_from_disk);
///fprintf(stderr, " -i ignore saved overlaps in *.ovlp*.bin files\n");
fprintf(stderr, " -N INT consider up to max(-D*coverage,-N) overlaps for each oriented read [%d]\n", asm_opt->max_n_chain);
fprintf(stderr, " -i ignore saved overlaps in *.ovlp* files\n");
fprintf(stderr, " -z INT length of adapters that should be removed [%d]\n", asm_opt->adapterLen);
fprintf(stderr, " -m INT size of popped large bubbles for contig graph [%lld]\n",
asm_opt->large_pop_bubble_size);
fprintf(stderr, " -p INT size of popped small bubbles for haplotype-resolved unitig graph [%lld]\n",
asm_opt->small_pop_bubble_size);
fprintf(stderr, " -m INT size of popped large bubbles for contig graph [%lld]\n", asm_opt->large_pop_bubble_size);
fprintf(stderr, " -p INT size of popped small bubbles for haplotype-resolved unitig graph [%lld]\n", asm_opt->small_pop_bubble_size);
fprintf(stderr, " -n INT small removed unitig threshold [%d]\n", asm_opt->max_short_tip);
fprintf(stderr, " -x FLOAT max overlap drop ratio [%.2g]\n", asm_opt->max_drop_rate);
fprintf(stderr, " -y FLOAT min overlap drop ratio [%.2g]\n", asm_opt->min_drop_rate);
@@ -51,8 +47,8 @@ void Print_H(hifiasm_opt_t* asm_opt)
fprintf(stderr, " -h show help information\n");
fprintf(stderr, " Trio-partition:\n");
fprintf(stderr, " -P FILE paternal trio index generated by \"yak count\" [NULL]\n");
fprintf(stderr, " -M FILE maternal trio index generated by \"yak count\" [NULL]\n");
fprintf(stderr, " -P FILE paternal trio index generated by \"yak count\" []\n");
fprintf(stderr, " -M FILE maternal trio index generated by \"yak count\" []\n");
fprintf(stderr, " -c INT lower bound of the binned k-mer's frequency [%d]\n", asm_opt->min_cnt);
fprintf(stderr, " -d INT upper bound of the binned k-mer's frequency [%d]\n", asm_opt->mid_cnt);
@@ -73,10 +69,10 @@ void init_opt(hifiasm_opt_t* asm_opt)
asm_opt->k_mer_length = 51;
asm_opt->mz_win = 51;
asm_opt->bf_shift = 37;
asm_opt->high_factor = 5.0f;
asm_opt->no_HPC = 0;
asm_opt->no_kmer_flt = 0;
asm_opt->max_n_chain = 400;
asm_opt->high_factor = 5.0f;
asm_opt->max_n_chain = 100;
asm_opt->k_mer_min_freq = 3;
asm_opt->k_mer_max_freq = 66;
asm_opt->load_index_from_disk = 1;
@@ -107,7 +103,7 @@ void destory_opt(hifiasm_opt_t* asm_opt)
}
}
void clear_opt(hifiasm_opt_t* asm_opt, int round)
void ha_opt_reset_to_round(hifiasm_opt_t* asm_opt, int round)
{
asm_opt->num_bases = 0;
asm_opt->num_corrected_bases = 0;
@@ -116,6 +112,14 @@ void clear_opt(hifiasm_opt_t* asm_opt, int round)
asm_opt->roundID = round;
}
void ha_opt_update_cov(hifiasm_opt_t *opt, int hom_cov)
{
int max_n_chain = (int)(hom_cov * opt->high_factor + .499);
if (opt->max_n_chain < max_n_chain)
opt->max_n_chain = max_n_chain;
fprintf(stderr, "[M::%s] updated max_n_chain to %d\n", __func__, opt->max_n_chain);
}
int check_file(char* name, const char* opt)
{
if(!name)
@@ -308,7 +312,7 @@ int CommandLine_process(int argc, char *argv[], hifiasm_opt_t* asm_opt)
int c;
while ((c = ketopt(&opt, argc, argv, 1, "hvt:o:k:lw:m:n:r:a:b:z:x:y:p:c:d:M:P:if:D:FN:", long_options)) >= 0) {
while ((c = ketopt(&opt, argc, argv, 1, "hvt:o:k:w:m:n:r:a:b:z:x:y:p:c:d:M:P:if:D:FN:", long_options)) >= 0) {
if (c == 'h')
{
Print_H(asm_opt);
@@ -325,7 +329,6 @@ int CommandLine_process(int argc, char *argv[], hifiasm_opt_t* asm_opt)
else if (c == 'r') asm_opt->number_of_round = atoi(opt.arg);
else if (c == 'k') asm_opt->k_mer_length = atoi(opt.arg);
else if (c == 'i') asm_opt->load_index_from_disk = 0;
else if (c == 'l') asm_opt->load_index_from_disk = 1;
else if (c == 'w') asm_opt->mz_win = atoi(opt.arg);
else if (c == 'D') asm_opt->high_factor = atof(opt.arg);
else if (c == 'F') asm_opt->no_kmer_flt = 1;