forked from aalbrechtsen/fastNGSadmix
-
Notifications
You must be signed in to change notification settings - Fork 4
/
fastNGSadmix.cpp
1976 lines (1726 loc) · 61.4 KB
/
fastNGSadmix.cpp
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
/*
Copyright (C) 2020 Emil Jorsboe - emil.jorsboe@bio.ku.dk
Kristian Hanghoj
Anders Albrechtsen
fastNGSadmix 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. For more see LICENSE file.
log:
g++ fastNGSadmix.cpp -lz -O3 -o fastNGSadmix
log: (with readplink function)
g++ fastNGSadmix.cpp readplinkV3.cpp -lz -O3 -o fastNGSadmix
debug:
g++ fastNGSadmix.cpp -lz -ggdb -O3 -o fastNGSadmix
debug: (with readplink function)
g++ fastNGSadmix.cpp readplinkV3.cpp -lz -ggdb -O3 -o fastNGSadmix
*/
#include <cstdio>
#include <cstdlib>
#include <cstring>
#include <ctime>
#include <cmath>
#include <limits>
#include <zlib.h>
#include <vector>
#include <pthread.h>
#include <signal.h>
#include <vector>
#include <sys/stat.h>
#include <map>
#include <iostream>
// stringpocalypse - still the best!
#include <string>
#include "readplinkV3.h"
//this is the max number of bytes perline
#define LENS 100000
//if we catch signal then quit program nicely
int SIG_COND =1;
double errTolMin=1e-5;
double errTolStart=0.05;
//frequencies and admixture coef cannot be less than this or more than 1-this
double errTol=errTolStart;
// checks if string is number
int validDouble(std::string someString){
int isNumber = 0;
int hasPoint = 0;
for(int i = 0; i<someString.length(); i++){
char s = someString[i];
if(s =='.'){
if(hasPoint){
isNumber = 0;
break;
}
hasPoint = 1;
isNumber = 1;
} else {
isNumber = (s=='0' or s=='1' or s=='2' or s=='3' or s=='4' or s=='5' or s=='6' or s=='7' or s=='8' or s=='9');
}
// leaves loop if any not numbers
if(isNumber==0){
break;
}
}
return(isNumber);
}
void minus1d(std::vector<double> &fst, std::vector<double> &sec,size_t x, std::vector<double> &res){
for(size_t i=0;i<x;i++){
res[i] = fst[i]-sec[i];
}
}
void minusFunc(std::vector< std::vector<double> > &fst,std::vector< std::vector<double> > &sec,size_t x,size_t y,std::vector< std::vector<double> > &res){
for(size_t i=0;i<x;i++){
for(size_t j=0;j<y;j++){
res[i][j] = fst[i][j]-sec[i][j];
}
}
}
double sumSquare(std::vector< std::vector<double> > &mat,size_t x,size_t y){
double tmp=0;
for(size_t i=0;i<x;i++){
for(size_t j=0;j<y;j++){
tmp += mat[i][j]*mat[i][j];
}
}
return tmp;
}
double sumSquare1d(std::vector<double > &mat,size_t x){
double tmp=0;
for(size_t i=0;i<x;i++){
tmp += mat[i]*mat[i];
}
return tmp;
}
double calcThres(std::vector<double> &d1, std::vector<double> &d2, int x){
// finds the largest difference between 2 arrays
// arrays has dimention x times y
double diff=fabs(d1[0]-d2[0]);
for(int i=1;i<x;i++){
if(fabs(d1[i]-d2[i])<diff){
diff=fabs(d1[i]-d2[i]);
}
}
return diff;
}
// function for keeping sure Q values do not become
// 0.0 as then division by 0 might occur, errTol is limit
void map2domainQ(std::vector<double> &Q, int nPop){
double sum=0;
for(int k=0;k<nPop;k++){
if(Q[k]<errTol){
Q[k] = errTol;
}
if(Q[k]>(1-errTol)){
Q[k] = 1-errTol;
}
sum+=Q[k];
}
for(int k=0;k<nPop;k++){
Q[k]=Q[k]/sum;
}
}
// function for keeping sure F values do not become
// 0.0 as then division by 0 might occur, errTol is limit
void map2domainF(std::vector< std::vector<double> > &F, int nSites, int nPop){
for(int s=0;s<nSites;s++)
for(int k=0;k<nPop;k++){
if(F[s][k]<errTol){
F[s][k] = errTol;
}
if(F[s][k]>1-errTol){
F[s][k] = 1-errTol;
}
}
}
////////////////////////
///@param str Filename given as a string.
int fexists(const char* str){
struct stat buffer ;
/// @return Function returns 1 if file exists.
return (stat(str, &buffer )==0 );
}
std::vector<std::string> dumpedFiles;
FILE *openFile(const char* a,const char* b){
if(0){
fprintf(stderr,"[%s] %s %s",__FUNCTION__,a,b);
}
std::string c1(a, strlen(a));
std::string c2(b, strlen(b));
std::string c = c1 + c2;
fprintf(stderr,"\t-> Dumping file: %s\n",c.c_str());
if(0&&fexists(c.c_str())){
fprintf(stderr,"File: %s exists will exist\n",c.c_str());
fflush(stderr);
exit(0);
}
dumpedFiles.push_back(c);
FILE *fp = fopen(c.c_str(),"w");
if (not(fp)){
fprintf(stderr,"File: %s cannot be created specify valid path\n",c.c_str());
fflush(stderr);
exit(0);
}
return fp;
}
gzFile openFileGz(const char* a,const char* b){
if(0){
fprintf(stderr,"[%s] %s %s",__FUNCTION__,a,b);
}
std::string c1(a, strlen(a));
std::string c2(b, strlen(b));
std::string c = c1 + c2;
fprintf(stderr,"\t-> Dumping file: %s\n",c.c_str());
if(0&&fexists(c.c_str())){
fprintf(stderr,"File: %s exists will exist\n",c.c_str());
fflush(stderr);
exit(0);
}
dumpedFiles.push_back(c);
gzFile fp = gzopen(c.c_str(),"w");
return fp;
}
// 0 indexed
double getGeno(const std::vector<double> &g, int row, int col){
if(col>2){
fprintf(stderr,"only one individaul in beagle file, thus only 3 columns \n");
exit(0);
}
return(g[3*row+col]);
}
double getFreq(const std::vector<double> &freq, int pops, int row, int col){
// vector is coded so first pops entries are row 0 (0pops...pops-1)
// then from pops...2*pops-1 is row 1 and so on
if(col>pops){
fprintf(stderr,"only %i pops in ref panel \n",pops);
exit(0);
}
return(freq[pops*row+col]);
}
//some struct with all the data from the beagle file
typedef struct{
std::vector<double> genos;
std::vector<char> major;
std::vector<char> minor;
// for snp ids, chr_pos
std::vector<std::string> id;
int nSites;
int nInd;
// map of ids in beagle file for finding overlap with ref
// id is like this: chr_pos
std::map <std::string,int> idMap;
}bgl;
// refPanel struct for reading in refPanel with header
typedef struct{
std::vector<std::string> id;
std::vector<int> chr;
std::vector<int> pos;
std::vector<std::string> name;
std::vector<char> A0;
std::vector<char> A1;
// change this to std::vector<double>, create function that can return freq
// has to know nSites and nPops
std::vector<double> freqs;
int refSites;
int pops;
std::vector<std::string> populations;
// has map of column to keep in ref for calculations
// is coded so key is old column number, from inputted ref
// value is new column (in ref for analysis) number + 1 (cause has to be above 0)
std::map <int,int> colsToKeep;
std::map <std::string,int> popsToKeep;
}refPanel;
// convert 0,1,2,3 to A,C,G,T if beagle coded thusly
char intToChar(char intLike){
if(intLike=='A' or intLike=='C' or intLike=='G' or intLike=='T'){
return(intLike);
} else {
if(intLike == '0'){
return('A');
} else if(intLike == '1'){
return('C');
} else if(intLike == '2'){
return('G');
} else if(intLike == '3'){
return('T');
} else{
fprintf(stderr,"Beagle nucleotide not valid must be A,C,G,T or 0,1,2,3\n");
exit(0);
}
}
}
/*
Returns the bgl struct containing all data from a beagle file.
It find the nsamples from counting the header
It finds the number of sites by queing every line in a std::vector
After the file has been read intotal it reloops over the lines in the vector and parses data
*/
bgl readBeagle(const char* fname, const std::map <std::string,int> &overlap) {
const char *delims = "\t \n";
gzFile fp = NULL;
// checking if file can be opened
if(Z_NULL==(fp=gzopen(fname,"r"))){
fprintf(stderr,"Error opening file: %s\n",fname);
exit(0);
}
bgl ret;
char buf[LENS];
gzgets(fp,buf,LENS);
strtok(buf,delims);
int ncols=1;
// reading first line in order to see nCol
while(strtok(NULL,delims)){
ncols++;
}
if(0!=(ncols-3) %3 ){
fprintf(stderr,"ncols=%d\n",ncols);
exit(0);
}
if(ncols > 6){
fprintf(stderr,"Only one individual in beagle file, looks like there are=%d\n",(ncols-3) / 3);
exit(0);
}
ret.nInd = (ncols-3)/3;
ret.nSites = overlap.size();
int refIndex = 0;
int bglIndex = 0;
std::string dummyID;
char dummyChar;
ret.id.assign(ret.nSites,dummyID);
ret.major.assign(ret.nSites,dummyChar);
ret.minor.assign(ret.nSites,dummyChar);
ret.genos.assign(ret.nSites*3,0);
while(gzgets(fp,buf,LENS)!=NULL){
// puts id of all sites in map for fast lookup
char* bglID = strtok(buf,delims);
std::string bglIDstring(bglID, strlen(bglID));
// because allele might be coded 0,1,2,3
char A0 = intToChar(strtok(NULL,delims)[0]);
char A1 = intToChar(strtok(NULL,delims)[0]);
if( tolower(A0) < tolower(A1) ){
bglIDstring = bglIDstring + "_" + A0 + "_" + A1;
} else{
bglIDstring = bglIDstring + "_" + A1 + "_" + A0;
}
// this keeps track of which position in beagle file a site is (+1 to be able to have .count() return TRUE)
ret.idMap[bglIDstring] = refIndex+1;
//then loop over the vector and parsing every line
if(overlap.count(bglIDstring)>0){
ret.id.at(refIndex)=bglIDstring;
ret.major.at(refIndex)=A0;
ret.minor.at(refIndex)=A1;
double tmpS = 0.0;
for(int i=0;i<ret.nInd*3;i++){
double gl = atof(strtok(NULL,delims));
ret.genos.at(refIndex*3+i)=gl;
if(gl<0){
fprintf(stderr,"Likelihoods must be positive\n");
fprintf(stderr,"site %d ind %d geno %d has value %f\n",bglIndex,int(i*1.0/3),i%3,getGeno(ret.genos,refIndex,i));
exit(0);
}
tmpS+=gl;
if(i==2 and !(tmpS>0)){
fprintf(stderr,"The sum of likelihoods for a genotypes must be positive\n");
fprintf(stderr,"individual %d site %d has sum %f\n",i,bglIndex,tmpS);
exit(0);
}
}
// counts which line of overlapping sites between bgl and ref panel
refIndex++;
}
// counts which line of beagle file
bglIndex++;
}
//clean up filepointer
gzclose(fp);
return ret;
}
// read in plink file and converts to a bgl struct (beagle file)
bgl readPlinkToBeagle(const char* plinkName, const std::map <std::string,int> &overlap) {
plink pl = readplink(plinkName);
if(pl.fam.individuals > 1){
fprintf(stderr,"More than one individual in input plink file - should only be one! \n");
exit(0);
}
bgl b;
b.nSites = overlap.size();
// can only have one individual is this program
b.nInd = 1;
int beagleIndex = 0;
char dummyChar;
std::string dummyID;
b.id.assign(b.nSites,dummyID);
b.major.assign(b.nSites,dummyChar);
b.minor.assign(b.nSites,dummyChar);
b.genos.assign(b.nSites*3,0);
for(int i=0;i<(pl.y);i++){
// bim id has chr_pos_A0_A1 ID (A0, A1 ordered alphabetically)
if(overlap.count(pl.bim.id[i])<1){
continue;
}
if(beagleIndex+1 > overlap.size()){
fprintf(stderr,"Duplicated sites - multiple sites with same position in plink file!! \n");
fprintf(stderr,"Use plink2 with --list-duplicate-vars suppress-first ids-only - and then --exclude \n");
exit(0);
}
if(pl.d[0][i]==0){
b.genos.at(beagleIndex*3)=0.0;
b.genos.at(beagleIndex*3+1)=0.0;
b.genos.at(beagleIndex*3+2)=1.0;
} else if(pl.d[0][i]==1){
b.genos.at(beagleIndex*3)=0.0;
b.genos.at(beagleIndex*3+1)=1.0;
b.genos.at(beagleIndex*3+2)=0.0;
} else if(pl.d[0][i]==2){
b.genos.at(beagleIndex*3)=1.0;
b.genos.at(beagleIndex*3+1)=0.0;
b.genos.at(beagleIndex*3+2)=0.0;
}
b.major.at(beagleIndex)=pl.bim.major[i];
b.minor.at(beagleIndex)=pl.bim.minor[i];
b.id.at(beagleIndex)=pl.bim.id[i];
// stores id of all overlapped sites from plink file,
// has index of which index in beagle file + 1
b.idMap[pl.bim.id[i]] = beagleIndex +1;
beagleIndex++;
}
kill_plink(pl);
return(b);
}
void readDouble(double **d,int x,int y,const char*fname,int neg){
fprintf(stderr,"opening : %s with x=%d y=%d\n",fname,x,y);
const char*delims=" \n\t";
FILE *fp = NULL;
if((fp=fopen(fname,"r"))==NULL){
fprintf(stderr,"cont open:%s\n",fname);
exit(0);
}
int lens=1000000;
char buf[lens];
for(int i=0;i<x;i++){
if(NULL==fgets(buf,lens,fp)){
fprintf(stderr,"Increase buffer\n");
exit(0);
}
if(neg)
d[i][0] = -atof(strtok(buf,delims));
else
d[i][0] = atof(strtok(buf,delims));
for(int j=1;j<y;j++){
if(neg)
d[i][j] = -atof(strtok(NULL,delims));
else
d[i][j] = atof(strtok(NULL,delims));
}
}
fclose(fp);
}
// function for reading in ref panel, has to have format id chr pos name A0_freq A1 pop1 pop2 ... (freq has to be of A0 allele)
refPanel readRefPanel(const char* fname, bgl b, const std::map <std::string,int> &includedPops, int nPop, const std::map <std::string,int> &overlap) {
const char *delims = "\t \n";
gzFile fp = NULL;
if(Z_NULL==(fp=gzopen(fname,"r"))){
fprintf(stderr,"Error opening file: %s\n",fname);
exit(0);
}
char buf[LENS];
refPanel ref;
int totalSites = 0;
int ncols = 0;
// keeps track of which new column (index = newCol-1) has to be above 0 for lookup in map
int newCol = 1;
//find number of columns
while(gzgets(fp,buf,LENS)!=NULL){
if(totalSites==0){
char* columnID = strtok(buf,delims);
while(columnID!=NULL){
ncols++;
// first 6 columns not freqs and has to at most K new columns included in ref
if(ncols>6){
// if in supplied populations or if no populations supplied include
if(includedPops.count(columnID)>0 or includedPops.empty()){
// prints out which populations chosen
fprintf(stderr,"Chosen pop %s\n",columnID);
std::string columnIDstring(columnID, strlen(columnID));
// for which columns to keep
ref.populations.push_back(columnIDstring);
// keep track of which new column it will be
ref.popsToKeep[columnIDstring] = newCol;
// so that can translate from org column (where 7th column is first freq column) to new column (that is +1 here for lookup purposes)
ref.colsToKeep[ncols-7] = newCol;
newCol++;
}
}
columnID = strtok(NULL,delims);
}
}
totalSites++;
}
gzclose(fp);
if(ncols<7){
// has to have at least 7 columns
fprintf(stderr,"Too few cols, ncols=%d\n",ncols);
exit(0);
}
ref.pops = ref.popsToKeep.size();
ref.refSites = b.nSites;
std::string dummyID;
char dummyChar;
ref.freqs.assign(ref.pops*ref.refSites,0);
ref.id.assign(ref.refSites,dummyID);
ref.chr.assign(ref.refSites,0);
ref.pos.assign(ref.refSites,0);
ref.name.assign(ref.refSites,dummyID);
// ref has A,C,G,T alleles
ref.A0.assign(ref.refSites,dummyChar);
ref.A1.assign(ref.refSites,dummyChar);
gzFile fp1 = NULL;
fp1=gzopen(fname,"r");
// for keeping track of which index in new ref with
int refIndex = 0;
int refSite = 0;
//keep track of how many columns are being read - should be the same as are in the header
int nColsRead = 0;
int line = 0;
while(gzgets(fp1,buf,LENS)!=NULL){
// looking at id value chr_pos for detecting overlap
char* id = strtok(buf,delims);
nColsRead++;
std::string stringID(id,strlen(id));
int refChr = atoi(strtok(NULL,delims));
nColsRead++;
int refPos = atoi(strtok(NULL,delims));
nColsRead++;
char* name = strtok(NULL,delims);
nColsRead++;
std::string stringName(name,strlen(name));
// ref has A,C,G,T alleles
char A0 = intToChar(strtok(NULL,delims)[0]);
nColsRead++;
char A1 = intToChar(strtok(NULL,delims)[0]);
nColsRead++;
// sorting also with alleles
if( tolower(A0) < tolower(A1) ){
stringID = stringID + "_" + A0 + "_" +A1;
} else{
stringID = stringID + "_" + A1 + "_" + A0;
}
// check if site is in overlap with beagle file
// otherwise continues to next site in ref
if(overlap.count(stringID) > 0){
// index of site in overlap - and thereby also in beagle file (only overlapped sites)
int overlapIndex = b.idMap[stringID]-1;
// the idMap of the beagle struct has index sites was placed on in beagle file (index+1)
ref.id.at(overlapIndex)=stringID;
ref.chr.at(overlapIndex)=refChr;
ref.pos.at(overlapIndex)=refPos;
ref.name.at(overlapIndex)=stringName;
// ref has A,C,G,T alleles
ref.A0.at(overlapIndex)=A0;
ref.A1.at(overlapIndex)=A1;
// reading in ref freqs
for(int i=0;i<(ncols-6);i++){
// check if org column to keep and thereby pop to keep in ref
if(ref.colsToKeep.count(i)>0){
// if bgl 1_1 A B GL(AA) GL(AB) GL(BB) Then ref 1 1 rs1 B A 1-f(B)
// minor is last allele in beagle file
if(ref.A0[overlapIndex]==b.minor[overlapIndex]){
// new col has to be - 1 for right index
// ref.freqs[refIndex][ref.colsToKeep[i]-1] = 1 - atof(strtok(NULL,delims));
ref.freqs.at(ref.pops*overlapIndex+(ref.colsToKeep[i]-1)) = 1 - atof(strtok(NULL,delims));
nColsRead++;
} else{
// ref.freqs[refIndex][ref.colsToKeep[i]-1] = atof(strtok(NULL,delims));
ref.freqs.at(ref.pops*overlapIndex+(ref.colsToKeep[i]-1)) = atof(strtok(NULL,delims));
nColsRead++;
}
if(getFreq(ref.freqs,ref.pops,overlapIndex,ref.colsToKeep[i]-1)<0){
fprintf(stderr,"Frequencies must be positive\n");
fprintf(stderr,"site %d, pop %d, has value %f\n",refSite,i,getFreq(ref.freqs,ref.pops,overlapIndex,ref.colsToKeep[i]-1));
exit(0);
}
} else{
// it has to skip cols that are not to be read in and move to next column which will be checked
strtok(NULL,delims);
nColsRead++;
}
}
refIndex++;
} else{
while(strtok(NULL,delims)!=NULL){
nColsRead++;
}
}
if(nColsRead!=ncols){
// it reads header so refSite actually refers to which sites and refSite + 1 line of file
fprintf(stderr,"Line of site %i (1-indexed) (row %i in refPanel_* file) is ill formated (probably does not have right number of columns)\n",refSite,refSite+1);
fprintf(stderr,"Line of site %i (1-indexed) has %i cols should have %i cols\n",refSite,nColsRead,ncols);
exit(0);
}
nColsRead = 0;
refSite++;
// only here if site was included in new ref
}
if(refIndex!=overlap.size()){
fprintf(stderr,"refIndex %i, overlap sites %lu\n",refIndex,overlap.size());
fprintf(stderr,"Seems like there are duplicate ids in input or reference panel!\n");
exit(0);
}
gzclose(fp1);
return ref;
}
void readDoubleGZ(double **d,int nSites,int nPop,const char*fname,int neg){
fprintf(stderr,"opening : %s with x=%d y=%d\n",fname,nSites,nPop);
const char*delims=" \n";
gzFile fp = NULL;
if((fp=gzopen(fname,"r"))==NULL){
fprintf(stderr,"cont open:%s\n",fname);
exit(0);
}
int lens=1000000;
char buf[lens];
std::vector<char*> tmp;
while(gzgets(fp,buf,LENS)){
tmp.push_back(strdup(buf));
}
// for reading in a bigger ref panel
// and getting intersecting sites
int freqSites = tmp.size();
fprintf(stderr,"This many ref sites: %i\n",freqSites);
for(int i=0;i<freqSites;i++){
if(NULL==gzgets(fp,buf,lens)){
fprintf(stderr,"Increase buffer\n");
exit(0);
}
if(neg){
d[i][0] = -atof(strtok(buf,delims));
}
else{
d[i][0] = atof(strtok(buf,delims));
}
for(int j=1;j<nPop;j++){
if(neg){
d[i][j] = -atof(strtok(NULL,delims));
}
else{
d[i][j] = atof(strtok(NULL,delims));
}
}
}
gzclose(fp);
}
// for reading number of individuals in each ref - nInd file
void readDouble1d(std::vector <double> &d,int nPop,const char*fname, std::map<std::string,int> popsToKeep){
fprintf(stderr,"Opening nInd file: %s with nPop=%d\n",fname,nPop);
const char*delims=" \n\t";
gzFile fp = NULL;
if((fp=gzopen(fname,"r"))==NULL){
fprintf(stderr,"cont open:%s\n",fname);
exit(0);
}
int lens=1000000 ;
char buf[lens];
d.assign(nPop,0);
std::vector<char*> tmp;
// keeps track of org index of pop to keep and new index of pop to keep
std::map<int,int> toKeep;
gzgets(fp,buf,LENS);
// looking at id value chr_pos for detecting overlap
char* word = strtok(buf,delims);
std::string stringID(word,strlen(word));
int orgCol = 0;
while(word!=NULL){
std::string stringWord(word,strlen(word));
if(popsToKeep.count(stringWord)>0){
// creates map of <nInd index, ref index> so can map from nInd order to ref order
// so if first element in nInd is second in ref
// the d array with nInd values has second element equal to nInd first value
toKeep[orgCol] = popsToKeep[stringWord];
}
word = strtok(NULL,delims);
orgCol++;
}
// first goes through the first line with names of pops, finds which should be included
// keeps track of which value at in nInd file, newCol has to be index+1, because has to be > 0 for lookup
if(toKeep.size()!=popsToKeep.size()){
fprintf(stderr,"nInd and ref panel do not have same size!\n");
exit(0);
}
int index = 0;
gzgets(fp,buf,LENS);
word = strtok(buf,delims);
while(word!=NULL){
if(toKeep.count(index)>0){
// because map index has to start at 1 for count method to work
d[toKeep[index]-1] = atof(word);
}
word = strtok(NULL,delims);
index++;
}
if(index!=orgCol){
fprintf(stderr,"nInd has different number of elements between row 1 and 2\n");
exit(0);
}
gzclose(fp);
}
void printDouble(const std::vector< std::vector<double> > &ret,size_t x,size_t y, int highestLike, int nConv, const std::vector<std::string> &populations ,FILE *fp){
for(size_t i=0;i<x;i++){
if(i==0){
for(size_t j=0;j<y;j++){
fprintf(fp,"%s ",populations[j].c_str());
}
fprintf(fp,"\n");
}
if(i<nConv and i == highestLike){
for(size_t j=0;j<y;j++){
fprintf(fp,"%.4f ",ret[i][j]);
}
fprintf(fp,"\n");
} else if(i>=nConv){
for(size_t j=0;j<y;j++){
fprintf(fp,"%.4f ",ret[i][j]);
}
fprintf(fp,"\n");
}
}
}
void printDoubleGz(const std::vector< std::vector<double> > &ret, size_t x, size_t y, const std::vector<std::string> &id, const std::vector<std::string> &populations ,gzFile fp){
for(size_t i=0;i<x;i++){
if(i==0){
gzprintf(fp,"id ");
for(size_t j=0;j<y;j++){
gzprintf(fp,"%s ",populations[j].c_str());
}
gzprintf(fp,"\n");
}
gzprintf(fp,"%s ",id[i].c_str());
for(size_t j=0;j<y;j++){
gzprintf(fp,"%.4f ",1-ret[i][j]);
}
gzprintf(fp,"\n");
}
}
// calculate log(likelihood) from likelihood function
double likelihood(const std::vector<double> &Q, const std::vector< std::vector<double> > &F,int nSites, int nPop, const std::vector<double> &genos, int ploidy){
double prod_ind = 0.0;
for(int j = 0; j < nSites; j++) {
double freq = 0.0;
for(int k = 0; k < nPop; k++) {
freq += (F[j][k])*Q[k];
}
// has to be like this, as I sort freqs
// prior to running this program
double f = freq;
// ploidy will be either 1 or 2!
if(ploidy==1){
double sum = getGeno(genos,j,0)*f;
sum += getGeno(genos,j,1)*(1-f);
prod_ind += log(sum);
} else if(ploidy==2){
double sum = getGeno(genos,j,0)* f * f;
sum += getGeno(genos,j,1)*2*f*(1-f);
sum += getGeno(genos,j,2)*(1-f)*(1-f);
prod_ind += log(sum);
}
}
return prod_ind;
}
// does bootstrapping sampling nSites random sites with replacement
void bootstrap(const std::vector<double> &genosOrg, std::vector<double> &genos, const std::vector< std::vector<double> > &F_orgOrg, std::vector< std::vector<double> > &F_org, std::vector< std::vector<double> > &F, int nPop, int nSites, int ploidy) {
for(int j=0;j<nSites;j++){
// generate random int from 0 to (nSites-1)
int row = std::rand() % nSites;
if(ploidy==1){
genos[3*j+0] = getGeno(genosOrg,row,0);
genos[3*j+1] = getGeno(genosOrg,row,1);
} else if(ploidy==2){
genos[3*j+0] = getGeno(genosOrg,row,0);
genos[3*j+1] = getGeno(genosOrg,row,1);
genos[3*j+2] = getGeno(genosOrg,row,2);
}
for(int k = 0; k < nPop; k++) {
F[j][k] = F_orgOrg[row][k];
F_org[j][k] = F_orgOrg[row][k];
}
}
}
// em algorithm not adjusting F at every step
void emUnadjusted(std::vector<double> &Q, std::vector< std::vector<double> > &F, int nSites, int nPop, const std::vector<double> &genos, std::vector<double> &Q_1, int ploidy) {
double sumAG[nPop];
double sumBG[nPop];
// makes sure neither F nor Q has 0 values
map2domainF(F, nSites, nPop);
map2domainQ(Q,nPop);
for(int k=0;k<nPop;k++){
sumAG[k]=0;
sumBG[k]=0;
}
for(int j=0;j<nSites;j++){
double fpart=0;
double fpartInv=0;
double expGG=0;
for(int k=0;k<nPop;k++){
// admixture adjusted freq, for each pop
fpart += F[j][k] * Q[k];
fpartInv += (1-F[j][k]) * Q[k];
if(ploidy==1){
double pp0=(1-fpart)*getGeno(genos,j,1);
double pp1=fpart*getGeno(genos,j,0);
double sum=pp0+pp1;
expGG =(pp1)/sum;
} else if(ploidy==2){
// pre GL (sites x 3) * (adjusted freq)
// for calculating H range 0-2, this is the expected genotype
double pp0=(1-fpart)*(1-fpart)*getGeno(genos,j,2);
double pp1=2*(1-fpart)*fpart* getGeno(genos,j,1);
double pp2=fpart*fpart* getGeno(genos,j,0);
double sum=pp0+pp1+pp2;
expGG =(pp1+2*pp2)/sum;
}
}
for(int k=0;k<nPop;k++){
sumAG[k] += expGG/(fpart) * (Q[k] * F[j][k]);
sumBG[k] += (ploidy-expGG)/fpartInv * (Q[k] * (1-F[j][k]));
}
}
for(int k=0;k<nPop;k++){
Q_1[k]=(sumAG[k] + sumBG[k])/(ploidy*nSites*1.0);
}
map2domainQ(Q_1,nPop);
}
// em algorithm adjusting F at every step
void em(std::vector<double> &Q, std::vector< std::vector<double> > &F, int nSites, const std::vector<double> &nInd, int nPop, const std::vector<double> &genos, std::vector< std::vector<double> > &F_1, std::vector<double> &Q_1, std::vector< std::vector<double> > &F_org, int ploidy) {
double sumAG[nPop];
double sumBG[nPop];
// makes sure neither F nor Q has 0 values
map2domainF(F, nSites, nPop);
map2domainF(F_org, nSites, nPop);
map2domainQ(Q,nPop);
double sumA[nPop];
double sumB[nPop];
for(int k=0;k<nPop;k++){
sumA[k]=0;
sumB[k]=0;
}
for(int j=0;j<nSites;j++){
for(int k=0;k<nPop;k++){
sumAG[k]=0;
sumBG[k]=0;
}
double fpart=0;
double fpartInv=0;
double expGG=0;
double sum=0;
for(int k=0;k<nPop;k++){
// admixture adjusted freq, for each pop
fpart += F[j][k] * Q[k];
fpartInv += (1-F[j][k]) * Q[k];
if(ploidy==1){
double pp0=(1-fpart)*getGeno(genos,j,1);
double pp1=fpart*getGeno(genos,j,0);
sum=pp0+pp1;
expGG = (pp1)/sum;
} else if(ploidy==2){
// pre GL (sites x 3) * (adjusted freq)
// for calculating H range 0-2, this is the expected genotype
double pp0=(1-fpart)*(1-fpart)*getGeno(genos,j,2);
double pp1=2*(1-fpart)*fpart* getGeno(genos,j,1);
double pp2=fpart*fpart* getGeno(genos,j,0);
sum=pp0+pp1+pp2;
expGG = (pp1+2*pp2)/sum;
}
}
for(int k=0;k<nPop;k++){
// similar to (H/(q*f))*q, for jth marker
sumAG[k] = (expGG) / (fpart) * (Q[k]*F[j][k]);
sumBG[k] = (ploidy-expGG) / fpartInv * (Q[k]*(1-F[j][k]));
sumA[k] += sumAG[k];
sumB[k] += sumBG[k];
sumAG[k] += nInd[k]*ploidy*F_org[j][k];
sumBG[k] += ploidy*nInd[k]-(ploidy*nInd[k]*F_org[j][k]);
}
for(int k=0;k<nPop;k++){