forked from ANGSD/angsd
-
Notifications
You must be signed in to change notification settings - Fork 0
/
abcAsso.cpp
3418 lines (2757 loc) · 91.2 KB
/
abcAsso.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
/*
emil [email protected] jan9 2019
has implemented new methods -doAsso 4, 5 and 6
thorfinn [email protected] dec17 2012
has modified so no sites[].chromo etc are used
anders [email protected] made this.
logistic regression changed 612new/613
part of angsd
*/
#include <cmath>
#include <htslib/kstring.h>
#include "shared.h"
#include "analysisFunction.h"
#include "abcFreq.h"
#include "abcAsso.h"
void abcAsso::printArg(FILE *argFile){
fprintf(argFile,"-------------\n%s:\n",__FILE__);
fprintf(argFile,"\t-doAsso\t%d\n",doAsso);
fprintf(argFile,"\t1: Frequency Test (Known Major and Minor)\n");
fprintf(argFile,"\t2: Score Test\n");
fprintf(argFile,"\t4: Latent genotype model\n");
fprintf(argFile,"\t5: Score Test with latent genotype model - hybrid test\n");
fprintf(argFile,"\t6: Dosage regression\n");
fprintf(argFile,"\t7: Latent genotype model (wald test)\n");
fprintf(argFile," Frequency Test Options:\n");
fprintf(argFile,"\t-yBin\t\t%s\t(File containing disease status)\t\n\n",yfile1);
fprintf(argFile," Score, Latent, Hybrid and Dosage Test Options:\n");
fprintf(argFile,"\t-yBin\t\t%s\t(File containing disease status)\n",yfile1);
fprintf(argFile,"\t-yCount\t\t%s\t(File containing count phenotypes)\n",yfile2);
fprintf(argFile,"\t-yQuant\t\t%s\t(File containing phenotypes)\n",yfile3);
fprintf(argFile,"\t-cov\t\t%s\t(File containing additional covariates)\n",covfile);
fprintf(argFile,"\t-model\t%d\n",model);
fprintf(argFile,"\t1: Additive/Log-Additive (Default)\n");
fprintf(argFile,"\t2: Dominant\n");
fprintf(argFile,"\t3: Recessive\n\n");
fprintf(argFile,"\t-minHigh\t%d\t(Require atleast minHigh number of high credible genotypes)\n",minHigh);
fprintf(argFile,"\t-minCount\t%d\t(Require this number of minor alleles, estimated from MAF)\n",minCount);
fprintf(argFile,"\t-assoThres\t%f\tThreshold for logistic regression\n",assoThres);
fprintf(argFile,"\t-assoIter\t%d\tNumber of iterations for logistic regression\n",assoIter);
fprintf(argFile,"\t-emThres\t%f\tThreshold for convergence of EM algorithm in doAsso 4 and 5\n",emThres);
fprintf(argFile,"\t-emIter\t%d\tNumber of max iterations for EM algorithm in doAsso 4 and 5\n\n",emIter);
fprintf(argFile,"\t-doPriming\t%d\tPrime EM algorithm with dosage derived coefficients (0: no, 1: yes - default) \n\n",doPriming);
fprintf(argFile," Hybrid Test Options:\n");
fprintf(argFile,"\t-hybridThres\t\t%f\t(p-value value threshold for when to perform latent genotype model)\n",hybridThres);
fprintf(argFile,"Examples:\n\tPerform Frequency Test\n\t \'./angsd -yBin pheno.ybin -doAsso 1 -GL 1 -out out -doMajorMinor 1 -minLRT 24 -doMaf 2 -doSNP 1 -bam bam.filelist'\n");
fprintf(argFile,"\tPerform Score Test\n\t \'./angsd -yBin pheno.ybin -doAsso 2 -GL 1 -doPost 1 -out out -doMajorMinor 1 -minLRT 24 -doMaf 2 -doSNP 1 -bam bam.filelist'\n");
fprintf(argFile,"\n");
}
void abcAsso::getOptions(argStruct *arguments){
doAsso=angsd::getArg("-doAsso",doAsso,arguments);
if(doAsso==3){
fprintf(stderr,"\t-> -doAsso 3 is deprecated from version 0.615 \n");
exit(0);
}
if(doAsso==0)
return;
doMaf=angsd::getArg("-doMaf",doMaf,arguments);
adjust=angsd::getArg("-adjust",adjust,arguments);
model=angsd::getArg("-model",model,arguments);
minCov=angsd::getArg("-minCov",minCov,arguments);
dynCov=angsd::getArg("-dynCov",dynCov,arguments);
minHigh=angsd::getArg("-minHigh",minHigh,arguments);
doPrint=angsd::getArg("-doPrint",doPrint,arguments);
minCount=angsd::getArg("-minCount",minCount,arguments);
sitePerm=angsd::getArg("-sitePerm",sitePerm,arguments);
GL=angsd::getArg("-GL",GL,arguments);
covfile=angsd::getArg("-cov",covfile,arguments);
doPost=angsd::getArg("-doPost",doPost,arguments);
hybridThres=angsd::getArg("-hybridThres",hybridThres,arguments);
assoThres=angsd::getArg("-assoThres",assoThres,arguments);
assoIter=angsd::getArg("-assoIter",assoIter,arguments);
emThres=angsd::getArg("-emThres",emThres,arguments);
emIter=angsd::getArg("-emIter",emIter,arguments);
doPriming=angsd::getArg("-doPriming",doPriming,arguments);
yfile1=angsd::getArg("-yBin",yfile1,arguments);
if(yfile1!=NULL)
isBinary=1;
yfile2=angsd::getArg("-yCount",yfile2,arguments);
if(yfile2!=NULL)
isCount=1;
yfile3=angsd::getArg("-yQuant",yfile3,arguments);
if(yfile3!=NULL)
isQuant=1;
if(doPrint)
fprintf(stderr,"finished [%s]\t[%s]\n",__FILE__,__FUNCTION__);
if(doPrint)
fprintf(stderr,"staring [%s]\t[%s]\n",__FILE__,__FUNCTION__);
if(doAsso && doMaf==0){
fprintf(stderr,"Error: you must estimate the maf (-doMaf) in order to perform association \n");
exit(0);
}
if(doAsso && yfile1==NULL && yfile2==NULL && yfile3==NULL){
fprintf(stderr,"Error: you must provide a phenotype file (-yBin or -yQuant) to perform association \n");
exit(0);
}
if(doAsso && arguments->inputtype==INPUT_BEAGLE&&doAsso==1){
fprintf(stderr,"Error: Only doAsso=2 can be performed on posterior input\n");
exit(0);
}
if(doAsso && arguments->inputtype!=INPUT_BEAGLE&&(doAsso==2)&&doPost==0){
fprintf(stderr,"Error: For doAsso=2 you must estimate the posterior probabilites for the genotypes (-doPost 1) \n");
exit(0);
}
if((isBinary+isCount+isQuant)>1){
fprintf(stderr,"Error: Only one of -yBin, yQuant and -yCount \n");
exit(0);
}
}
abcAsso::abcAsso(const char *outfiles,argStruct *arguments,int inputtype){
multiOutfile=NULL;
bufstr.s=NULL;bufstr.l=bufstr.m=0;
model=1;
doPrint=0;
doAsso=0;
GL=0;
doPost=0;
isBinary=0;
isCount=0;
isQuant=0;
sitePerm=0;//not for users
covfile=NULL;
yfile1=NULL;
yfile2=NULL;
yfile3=NULL;
minHigh=10;
minCount=10;
dynCov=0;//not for users
minCov=5;//not for users
adjust=1;//not for users
doMaf=0;
assoIter=100;
assoThres=1e-06;
emIter=40;
emThres=1e-04;
hybridThres=0.05;
doPriming=1;
//from command line
if(arguments->argc==2){
if(!strcasecmp(arguments->argv[1],"-doAsso")){
printArg(stdout);
exit(0);
}else
return;
}
getOptions(arguments);
if(doAsso==0){
shouldRun[index] =0;
return;
}
printArg(arguments->argumentFile);
//read phenotype - binary or count (poisson)
if(isBinary)
ymat = angsd::getMatrix(yfile1,1,100000);
else if(isCount)
ymat = angsd::getMatrix(yfile2,1,100000);
else
ymat = angsd::getMatrix(yfile3,0,100000);
//read covariates
if(covfile!=NULL)
covmat = angsd::getMatrix(covfile,0,100000);
else{
covmat.x=0;
covmat.y=0;
covmat.matrix=NULL;
}
if(covfile!=NULL&&(covmat.x!=ymat.x)){
fprintf(stderr,"The number of covariates (%d) does not match the number of phenotypes (%d)\n",covmat.x,ymat.x);
exit(0);
}
if(!isBinary && doAsso==1){
fprintf(stderr,"Error: Only doAsso=2 can be performed on quantitative traits\n");
exit(0);
}
// check cov and ymat
check_pars(covmat,ymat,isBinary);
//make output files
multiOutfile = new BGZF*[ymat.y];
const char* postfix;
postfix=".lrt";
if(covfile!=NULL&&minCov>0){
int keepList[ymat.x];
for(int i=0 ; i < ymat.x;i++) {
keepList[i]=1;
for(int yi=0;yi<ymat.y;yi++) {
if(ymat.matrix[i][yi]==-999)
keepList[i]=0;
}
for(int ci=0;ci<covmat.y;ci++) {
if(covmat.matrix[i][ci]==-999)
keepList[i]=0;
}
}
int nCov=0;
int count[covmat.y];
for(int ci=0;ci<covmat.y;ci++) {
count[ci]=0;
for(int i=0 ; i < ymat.x;i++) {
if(keepList[i]==0)
continue;
if(covmat.matrix[i][ci]!=0){
count[ci]++;
}
}
if(count[ci]<minCov){
fprintf(stderr,"Error: Cov #%d only has %d non zero entries\n",ci,count[ci]);
}
else
nCov++;
}
if(!dynCov&&covmat.y!=nCov){
fprintf(stderr,"Error: Creating new covariant matrix with %d columns\n",nCov);
exit(0);
}
else if(covmat.y!=nCov){
// angsd::printMatrix(covmat,stderr);
fprintf(stderr,"Error: Creating new covariant matrix with %d columns\n",nCov);
angsd::Matrix<double> newmat;
newmat.x=covmat.x;
newmat.y=nCov;
newmat.matrix=new double*[covmat.x];
for(int xi=0;xi<covmat.x;xi++){
newmat.matrix[xi] = new double[nCov];
int tempCount=0;
for(int ci=0;ci<covmat.y;ci++){
if(count[ci]>minCov){
newmat.matrix[xi][tempCount]=covmat.matrix[xi][ci];
tempCount++;
}
}
}
angsd::deleteMatrix(covmat);
covmat=newmat;
}
}
//open outfiles
for(int i=0;i<ymat.y;i++){
char ary[5000];
snprintf(ary,5000,"%s%d.gz",postfix,i);
multiOutfile[i] = NULL;
multiOutfile[i] = aio::openFileBG(outfiles,ary);
}
//print header
bufstr.l=0;
// hybrid model score model first, then if significant use EM asso
if(doAsso==5){
ksprintf(&bufstr,"Chromosome\tPosition\tMajor\tMinor\tFrequency\tN\tLRTscore\thigh_WT/HE/HO\tLRTem\tbeta\tSE\temIter\n");
//EM asso or dosage model
} else if(doAsso==4 or doAsso==7){
ksprintf(&bufstr,"Chromosome\tPosition\tMajor\tMinor\tFrequency\tN\tLRT\tbeta\tSE\thigh_WT/HE/HO\temIter\n");
} else if(doAsso==6){
ksprintf(&bufstr,"Chromosome\tPosition\tMajor\tMinor\tFrequency\tN\tLRT\tbeta\tSE\thigh_WT/HE/HO\n");
} else if(doAsso==2)
ksprintf(&bufstr,"Chromosome\tPosition\tMajor\tMinor\tFrequency\tN\tLRT\thigh_WT/HE/HO\n");
else
ksprintf(&bufstr,"Chromosome\tPosition\tMajor\tMinor\tFrequency\tLRT\n");
for(int yi=0;yi<ymat.y;yi++)
aio::bgzf_write(multiOutfile[yi],bufstr.s,bufstr.l);
bufstr.l=0;
}
// desctructor
abcAsso::~abcAsso(){
if(doPrint)
fprintf(stderr,"staring [%s]\t[%s]\n",__FILE__,__FUNCTION__);
if(doAsso==0)
return;
for(int i=0;i<ymat.y;i++)
if(multiOutfile[i]!=NULL)
bgzf_close(multiOutfile[i]);
delete [] multiOutfile;
if(covfile!=NULL)
angsd::deleteMatrix(covmat);
angsd::deleteMatrix(ymat);
}
void abcAsso::clean(funkyPars *pars){
if(doAsso==0)
return;
assoStruct *assoc =(assoStruct*) pars->extras[index];
if(assoc->stat!=NULL)
for(int yi=0;yi<ymat.y;yi++)
delete[] assoc->stat[yi];
delete[] assoc->stat;
if(doAsso==5){
delete[] assoc->highWt;
delete[] assoc->highHe;
delete[] assoc->highHo;
delete[] assoc->betas;
delete[] assoc->SEs;
for(int yi=0;yi<ymat.y;yi++)
delete[] assoc->statOther[yi];
delete[] assoc->statOther;
} else if(doAsso==2){
delete[] assoc->highWt;
delete[] assoc->highHe;
delete[] assoc->highHo;
} else if(doAsso==4 or doAsso==6 or doAsso==7){
delete[] assoc->highWt;
delete[] assoc->highHe;
delete[] assoc->highHo;
delete[] assoc->betas;
delete[] assoc->SEs;
}
if(doAsso==4 or doAsso==5 or doAsso==7){
delete[] assoc->emIter;
}
if(assoc->keepInd!=NULL)
for( int yi =0;yi<ymat.y;yi++)
delete[] assoc->keepInd[yi];
delete[] assoc->keepInd;
delete assoc;
}
void abcAsso::print(funkyPars *pars){
if(doAsso==0)
return;
if(doPrint)
fprintf(stderr,"staring [%s]\t[%s]\n",__FILE__,__FUNCTION__);
printDoAsso(pars);
}
assoStruct *allocAssoStruct(){
assoStruct *assoc = new assoStruct;
assoc->SEs=NULL;
assoc->betas=NULL;
assoc->stat=NULL;
assoc->statOther=NULL;
assoc->keepInd=NULL;
assoc->highWt = NULL;
assoc->highHe = NULL;
assoc->highHo = NULL;
assoc->emIter = NULL;
return assoc;
}
void abcAsso::run(funkyPars *pars){
if(doAsso==0)
return;
assoStruct *assoc = allocAssoStruct();
pars->extras[index] = assoc;
if(doAsso==2 or doAsso==4 or doAsso==5 or doAsso==6 or doAsso==7){
assoc->highWt=new int[pars->numSites];
assoc->highHe=new int[pars->numSites];
assoc->highHo=new int[pars->numSites];
}
if(doAsso==1){
frequencyAsso(pars,assoc);
} else if(doAsso==2){
scoreAsso(pars,assoc);
} else if(doAsso==4){
assoc->betas=new double[pars->numSites];
assoc->SEs=new double[pars->numSites];
assoc->emIter=new int[pars->numSites];
emAsso(pars,assoc);
} else if(doAsso==6){
assoc->betas=new double[pars->numSites];
assoc->SEs=new double[pars->numSites];
dosageAsso(pars,assoc);
} else if(doAsso==5){
assoc->betas=new double[pars->numSites];
assoc->SEs=new double[pars->numSites];
assoc->emIter=new int[pars->numSites];
hybridAsso(pars,assoc);
} else if(doAsso==7){
assoc->betas=new double[pars->numSites];
assoc->SEs=new double[pars->numSites];
assoc->emIter=new int[pars->numSites];
emAssoWald(pars,assoc);
}
}
void abcAsso::check_pars(angsd::Matrix<double> &cov, angsd::Matrix<double> &phe, int isBinary){
// check if all values of first column in cov file is 1
int interceptChecker = 0;
// check if phenotype 0 or 1 - for quantitative
int isBinaryQuan = 1;
for(int j=0;j<phe.y;j++){
for(int i=0;i<phe.x;i++){
//if logistic regression check if phenotypes are 0 or 1
if(isBinary){
isBinaryQuan = 0;
if(phe.matrix[i][j]!=0 and phe.matrix[i][j]!=1 and phe.matrix[i][j]!=-999){
fprintf(stderr,"Phenotypes are not binary (0 or 1) for logistic model!\n");
exit(1);
}
} else if(isCount){
isBinaryQuan = 0;
if(ceil(phe.matrix[i][j])!=floor(phe.matrix[i][j])){
fprintf(stderr,"Phenotypes are not count data (must be integer) for poisson model!\n");
exit(1);
}
} else{
if(phe.matrix[i][j]!=0 and phe.matrix[i][j]!=1){
isBinaryQuan = 0;
}
}
}
if(isBinaryQuan){
fprintf(stderr,"\n");
fprintf(stderr,"#######################################################################################\n");
fprintf(stderr,"#######################################################################################\n");
fprintf(stderr,"WARNING: Phenotype nr. %i for linear model appears to be binary, run a logistic model instead!\n",j);
fprintf(stderr,"#######################################################################################\n");
fprintf(stderr,"#######################################################################################\n");
fprintf(stderr,"\n");
}
isBinaryQuan=1;
}
for(int j=0;j<cov.y;j++){
for(int i=0;i<cov.x;i++){
if(cov.matrix[i][j]==1){
interceptChecker++;
}
}
// p->len is equal to number of indis in plink file
// if these are same length, means user has column of 1s in .cov file
if(interceptChecker==cov.x){
fprintf(stderr,"Column of 1s for intercept should not be included in covariate,\n");
fprintf(stderr,"as this will be added automatically.\n");
exit(1);
}
interceptChecker=0;
}
}
void abcAsso::frequencyAsso(funkyPars *pars,assoStruct *assoc){
if(doPrint)
fprintf(stderr,"staring [%s]\t[%s]\n",__FILE__,__FUNCTION__);
if(pars->nInd!=ymat.x){
fprintf(stderr,"The number of sequenced individuals (%d) does not match the number of phenotypes (%d)\n",pars->nInd,ymat.x);
fflush(stderr);
exit(0);
}
if(ymat.y!=1){
fprintf(stderr,"Only one phenotype allowed for -doAsso 1\n");
fflush(stderr);
exit(0);
}
double **stat = new double*[ymat.y];
for(int yi=0;yi<ymat.y;yi++)
stat[yi] = new double[pars->numSites];
int y[pars->nInd];
for(int i=0;i<pars->nInd;i++)
y[i]=ymat.matrix[i][0];
double **like0;//genotype likelihood for controls
double **like1;//genotype likelihood for cases
double **likeAll;//genotype likelihood for cases and controls
int Ncases=0; //number of cases
int Ncontrols=0; //number of cases
int Nall=0; //number of cases and controls
int cases[pars->nInd];
int controls[pars->nInd];
int all[pars->nInd];
for(int i=0;i<pars->nInd;i++){
cases[i]=0;
controls[i]=0;
all[i]=0;
if((int)y[i]==1){
Ncases++;
cases[i]=1;
Nall++;
all[i]=1;
}
if((int)y[i]==0){
Ncontrols++;
controls[i]=1;
Nall++;
all[i]=1;
}
if(doPrint)
fprintf(stderr,"all, case, control: %d %d %d\n",all[i],cases[i],controls[i]);
}
if(doPrint)
fprintf(stderr,"count complete [%s]\t[%s]\n",__FILE__,__FUNCTION__);
if(doAsso==1){
like0=angsd::get3likesRMlow(pars,controls);
like1=angsd::get3likesRMlow(pars,cases);
likeAll=angsd::get3likesRMlow(pars,all);
}
if(doPrint)
fprintf(stderr,"like complete [%s]\t[%s]\n",__FILE__,__FUNCTION__);
for(int s=0;s<pars->numSites;s++){//loop overs sites
stat[0][s]=-999;
if(pars->keepSites[s]==0)
continue;
if(doAsso==1){
if(doPrint)
fprintf(stderr,"do freq [%s]\t[%s]\n",__FILE__,__FUNCTION__);
double score0=abcFreq::likeFixedMinor(abcFreq::emFrequency_ext(like0[s],Ncontrols,NULL,Ncontrols),like0[s],Ncontrols);
//likelihood for the cases
double score1=abcFreq::likeFixedMinor(abcFreq::emFrequency_ext(like1[s],Ncases,NULL,Ncases),like1[s],Ncases);
//likelihood for all individuals
double scoreNull=abcFreq::likeFixedMinor(abcFreq::emFrequency_ext(likeAll[s],Nall,NULL,Nall),likeAll[s],Nall);
//likelhood ratio statistics \sim chi^2
double LRT=-2*(score0+score1-scoreNull);
stat[0][s]=LRT;
}
}
assoc->stat=stat;
for(int s=0;s<pars->numSites;s++){
delete[] like0[s];
delete[] like1[s];
delete[] likeAll[s];
}
delete[] like0;
delete[] like1;
delete[] likeAll;
if(doPrint)
fprintf(stderr,"finish [%s]\t[%s]\n",__FILE__,__FUNCTION__);
}
void abcAsso::scoreAsso(funkyPars *pars,assoStruct *assoc){
if(doPrint)
fprintf(stderr,"staring [%s]\t[%s]\n",__FILE__,__FUNCTION__);
if(pars->nInd!=ymat.x){
fprintf(stderr,"The number of sequenced individuals (%d) does not match the number of phenotypes (%d)\n",pars->nInd,ymat.x);
exit(0);
}
int **keepInd = new int*[ymat.y];
double **stat = new double*[ymat.y];
for(int yi=0;yi<ymat.y;yi++){
stat[yi] = new double[pars->numSites];
keepInd[yi]= new int[pars->numSites];
}
for(int s=0;s<pars->numSites;s++){//loop overs sites
if(pars->keepSites[s]==0)
continue;
int *keepListAll = new int[pars->nInd];
for(int i=0 ; i<pars->nInd ;i++){
keepListAll[i]=1;
}
for(int yi=0;yi<ymat.y;yi++) { //loop over phenotypes
int *keepList = new int[pars->nInd];
keepInd[yi][s]=0;
for(int i=0 ; i<pars->nInd ;i++) {
keepList[i]=1;
if(keepListAll[i]==0||ymat.matrix[i][yi]==-999)
keepList[i]=0;
if(covfile!=NULL)
for(int ci=0;ci<covmat.y;ci++) {
if(covmat.matrix[i][ci]==-999)
keepList[i]=0;
}
if(keepList[i]==1)
keepInd[yi][s]++;
}
double *y = new double[pars->nInd];
for(int i=0 ; i<pars->nInd ;i++)
y[i]=ymat.matrix[i][yi];
freqStruct *freq = (freqStruct *) pars->extras[6];
stat[yi][s]=doAssociation(pars,pars->post[s],y,keepInd[yi][s],keepList,freq->freq[s],s,assoc);
//cleanup
delete [] y;
delete [] keepList;
} //phenotypes end
delete [] keepListAll;
} // sites end
assoc->stat=stat;
assoc->keepInd=keepInd;
}
///////////////////////////////////////////////////////////////////////////////
// INSERTED BY EMIL
///////////////////////////////////////////////////////////////////////////////
double abcAsso::standardError(double* start, angsd::Matrix<double> *design, angsd::Matrix<double> *postAll, double *y, double *post, int isBinary, int isCount,int *keepList, int nInd){
//problem is that not all y values might be included since some might have missing information
std::vector<double> ynew(design->x);
//only need individuals with valid phenotypic information
int count=0;
for(int i=0;i<nInd;i++){
if(keepList[i]){
ynew[count]=y[i];
count++;
}
}
if(postAll!=NULL){
double ret;
for(int i=0;i<design->x;i++){
double m = 0;
for(int j=0;j<design->y;j++){
m += design->matrix[i][j]*start[j];
}
// because y is indis long, each entry has to be used 3 times - design has 3*indis rows
size_t indexy = (size_t)floor(i/3);
if(isBinary){
double prob = exp(m)/(exp(m)+1.0);
// now handling if p is 0 or 1 (makes it very small or large)
m = angsd::bernoulli(ynew[indexy],prob,1);
} else if(isCount){
double lambda = exp(m);
m = angsd::poisson(ynew[indexy],lambda,1);
} else{
// density function of normal distribution
m = angsd::dnorm(ynew[indexy],m,start[design->y],1);
}
double tmp = m + log(post[i]);
postAll->matrix[(size_t)floor(i/3)][i % 3] = tmp;
}
// x is number of indis, y is 3
postAll->x = (size_t) design->x/3;
postAll->y = 3;
// to get rowSums
std::vector<double> postTmp(postAll->x);
for(int i=0;i<postAll->x;i++){
double tmp = 0;
double maxval = postAll->matrix[i][0];
for(int j=0;j<postAll->y;j++){
// find max - part of trick for doing log(p1+p2+p3)
maxval = std::max(maxval,postAll->matrix[i][j]);
}
// trick to avoid over/underflow - log(exp(log(val1)-log(max)) + ...) + log(max) = (exp(log(val1))/exp(log(max)))*(max) + ...
// same (exp(log(val1))/exp(log(max)))*(max)
postTmp[i] = log(exp(postAll->matrix[i][0]-maxval)+exp(postAll->matrix[i][1]-maxval)+exp(postAll->matrix[i][2]-maxval)) + maxval;
}
// divide each entry of a row with sum of that row
int df = postAll->x - design->y;
for(int i=0;i<postAll->x;i++){
double tmp = 0;
for(int j=0;j<postAll->y;j++){
postAll->matrix[i][j] -= postTmp[i];
}
}
//need to flatten the weights, which is p(s|y,G,phi,Q,f)
// so first four values is for first indi, and so on...
double weights[postAll->x*postAll->y];
int a = 0;
for(int i=0;i<postAll->x;i++)
for(int j=0;j<postAll->y;j++){
weights[a++] = exp(postAll->matrix[i][j]);
//check if issue with weights
if(exp(postAll->matrix[i][j])!=exp(postAll->matrix[i][j]) or std::isinf(exp(postAll->matrix[i][j]))){
fprintf(stderr,"Issue with weights being nan or inf\n");
return(-9);
}
}
std::vector<double> yTilde(design->x);
for(int i=0;i<design->x;i++)
for(int x=0;x<design->y;x++){
yTilde[i] += design->matrix[i][x]*start[x];
}
// has to be expected value - mu, therefore using link function exp/(1+exp)
if(isBinary){
for(int i=0;i<design->x;i++)
yTilde[i] = exp(yTilde[i]) / (1+exp(yTilde[i]));
}
// link function exp
if(isCount){
for(int i=0;i<design->x;i++)
yTilde[i] = exp(yTilde[i]);
}
int count=0;
// this has colSums for each col for each indi (3 rows)
std::vector<double> summedDesign((design->x/3)*design->y);
std::vector<double> tmp(design->x);
for(int i=0;i<design->x;i++){
// because y is indis long, each entry has to be used 3 times - design has 3*indis rows
size_t indexy = (size_t)floor(i/3);
// getting the residuals
if(isBinary || isCount){
tmp[i] = (ynew[indexy]-yTilde[i]) * weights[i];
} else{
tmp[i] = ((ynew[indexy]-yTilde[i]) / (start[design->y]*start[design->y]))*weights[i];
}
for(int x=0;x<design->y;x++){
if(i % 3 == 2){
// for each col take 3 rows for indis and take sum
summedDesign[count*design->y+x]=design->matrix[i-2][x]*tmp[i-2]+design->matrix[i-1][x]*tmp[i-1]+design->matrix[i][x]*tmp[i];
}
}
if(i % 3 == 2){
count++;
}
}
double invSummedDesign[design->y*design->y];
for(int i=0;i<design->y*design->y;i++)
invSummedDesign[i]=0;
// this is doing the matrix product of (X)^T*W*X
for(int x=0;x<design->y;x++)
for(int y=0;y<design->y;y++)
for(int i=0;i<(int)floor(design->x/3);i++){
invSummedDesign[x*design->y+y]+=summedDesign[i*design->y+x]*summedDesign[i*design->y+y];
}
// doing inv((X)^T*W*X)
int singular=angsd::svd_inverse(invSummedDesign, design->y, design->y);
return(sqrt(invSummedDesign[0]));
} else{
std::vector<double> yTilde(design->x);
for(int i=0;i<design->x;i++)
for(int x=0;x<design->y;x++){
yTilde[i] += design->matrix[i][x]*start[x];
}
// has to be expected value - mu, therefore using link function exp/(1+exp)
if(isBinary){
for(int i=0;i<design->x;i++)
yTilde[i] = exp(yTilde[i]) / (1+exp(yTilde[i]));
}
// link function exp
if(isCount){
for(int i=0;i<design->x;i++)
yTilde[i] = exp(yTilde[i]);
}
std::vector<double> summedDesign(design->x*design->y);
std::vector<double> tmp(design->x);
//design has 3*indis rows
for(int i=0;i<design->x;i++){
//ynew is indis long, so this is to be able to use that each value 3 times
size_t indexy = (size_t)floor(i/3);
if(isBinary || isCount){
tmp[i] = (ynew[indexy]-yTilde[i]);
} else{
tmp[i] = ((ynew[indexy]-yTilde[i]) / (start[design->y]*start[design->y]));
}
for(int x=0;x<design->y;x++){
summedDesign[i*design->y+x]=design->matrix[i][x]*tmp[i];
}
}
double invSummedDesign[design->y*design->y];
for(int i=0;i<design->y*design->y;i++)
invSummedDesign[i]=0;
// this is doing the matrix product of (X)^T*W*X
for(int x=0;x<design->y;x++)
for(int y=0;y<design->y;y++)
for(int i=0;i<design->x;i++){
invSummedDesign[x*design->y+y]+=summedDesign[i*design->y+x]*summedDesign[i*design->y+y];
}
// doing inv((X)^T*W*X)
int singular=angsd::svd_inverse(invSummedDesign, design->y, design->y);
return(sqrt(invSummedDesign[0]));
}
}
double abcAsso::dosageAssoc(funkyPars *p,angsd::Matrix<double> *design,angsd::Matrix<double> *designNull,double *postOrg,double *yOrg,int keepInd,int *keepList,double freq,int s,assoStruct *assoc,int model, int isBinary, int isCount, double* start, int fullModel){
int maf0=1;
// if too rare do not run analysis
if(freq>0.995||freq<0.005){
maf0=0;
}
int highWT=0;
int highHE=0;
int highHO=0;
double covMatrix[(covmat.y+2)*keepInd];
double covMatrixNull[(covmat.y+1)*keepInd];
double yfit[keepInd];
double yNull[keepInd];
int count = 0;
double post[keepInd*3];
designNull->x=keepInd;
//covars + intercept
designNull->y=covmat.y+1;
// WLS (weighted) model with genotypes
for(int i=0;i<p->nInd;i++){
if(keepList[i]){
yNull[count]=yOrg[i];
post[count*3+0]=postOrg[i*3+0];
post[count*3+1]=postOrg[i*3+1];
post[count*3+2]=postOrg[i*3+2];
covMatrixNull[count] = 1;
designNull->matrix[count][0] = 1;
for(int c=0;c<covmat.y;c++){
designNull->matrix[count][c+1] = covmat.matrix[i][c];
covMatrixNull[(c+1)*keepInd+count] = covmat.matrix[i][c];
}
count++;
}
}
// need intercept
double startNull[covmat.y+2];
for(int i=0;i<(covmat.y+2);i++){
startNull[i] = drand48()*2-1;
}
if(count!=keepInd){
fprintf(stderr,"[%s] wrong number of non missing\n",__FUNCTION__);
fflush(stderr);
exit(0);
}
if(isBinary){
getFitBin(yfit,yNull,covMatrixNull,keepInd,designNull->y,startNull);
} else if(isCount){
getFitPois(yfit,yNull,covMatrixNull,keepInd,designNull->y,startNull);
} else{
getFit(yfit,yNull,covMatrixNull,keepInd,designNull->y,startNull);
}
double llhNull = logLike(startNull,yNull,designNull,post,isBinary,isCount,0);
start[0] = drand48()*2-1;
double y[keepInd];
count=0;
design->x=keepInd;
design->y=covmat.y+2;
// WLS (weighted) model with genotypes
for(int i=0;i<p->nInd;i++){
if(keepList[i]){
y[count]=yOrg[i];
// 1: add, 2: dom, 3: rec
if(model==3){
design->matrix[count][0] = post[count*3+2];
covMatrix[count] = post[count*3+2];
} else if(model==2){
design->matrix[count][0] = post[count*3+2] + post[count*3+1];
covMatrix[count] = post[count*3+2] + post[count*3+1];
} else{
// getting dosage of the minor allele or second allele of beagle file
design->matrix[count][0] = 2*post[count*3+2] + post[count*3+1];
covMatrix[count] = 2*post[count*3+2] + post[count*3+1];
}
//if rec model, WT+HE->WT, HO->HE
if(model==3){
if((post[count*3+0]+post[count*3+1])>0.90)
highWT++;
if(post[count*3+2]>0.90)
highHE++;
} else if(model==2){
//if dom model, WT->WT, HE+HO->HE, HO->0
if(post[count*3+0]>0.90)
highWT++;
if((post[count*3+1]+post[count*3+2])>0.90)
highHE++;
} else{
if(post[count*3+0]>0.90)
highWT++;
if(post[count*3+1]>0.90)
highHE++;
if(post[count*3+2]>0.90)
highHO++;
}