-
Notifications
You must be signed in to change notification settings - Fork 8
/
pleiotropy_analysis.m
293 lines (243 loc) · 12.1 KB
/
pleiotropy_analysis.m
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
% (c) 2019 University of Oslo
% Pleiotropy-informed conditional and conjunctional false discovery rate
% Running
% - pleioFDR : Conditional / Conjunction FDR
% - plotqq : Q-Q plots
% - plot_enrichment : Fold Enrichment
% - plot_manhattan : Manhattan plots
%
% Requires the following data to be in your workspace:
%
% options : an instance of pleioOpt class with options and settings for this run
% traitfile1 : string, .mat file containing main phenotype
% traitfiles : cell array of strings, .mat files for secondary phenotypes
% traitname1 : string, name of the main phenotype
% traitnames : cell array of strings, name of secondary phenotypes
% LDmat : logical sparse matrix, nsnp x nsnp, LD matrix cut at certain threshold
% chrnumvec : matrix, nsnp x 1, chromosome (1..22)
% posvec : matrix, nsnp x 1, base pair position on chromosome
% ivec0 : logical matrix, nsnp x 1, indicating variants to use for genomic correction
% (e.g. intergenic)
%% LOAD OCTAVE PACKAGES
if is_octave()
pkg load statistics;
pkg load nan;
graphics_toolkit('gnuplot');
end
%% INPUT VALIDATION
if ~exist('options', 'var'), options = pleioOpt(); end
if ~isempty(options.outputdir), outputdir = options.outputdir; end
if ~exist('outputdir', 'var') || isempty(outputdir), outputdir='test'; end
if exist('fishercomb', 'var') && fishercomb ~= options.fishercomb
error('fishercomb must be specified via pleioOpt options');
end
if exist('correct_for_sample_overlap', 'var') && ...
correct_for_sample_overlap ~= options.correct_for_sample_overlap
error('correct_for_sample_overlap must be specified via pleioOpt options');
end
%% START DIARY
mkdir(outputdir)
diary(sprintf('%s/%s%s_%s_%g.log', outputdir,traitname1,...
sprintf('_%s',traitnames{:}), options.stattype, options.fdrthresh))
display(options)
%% LOAD FILES
fprintf('Loading GWAS .mat files... ')
[logpvec1,zvec1]=load_gwas(traitfile1, length(chrnumvec), options.dummy_zscore);
[logpmat2,zmat2]=load_gwas(traitfiles, length(chrnumvec), options.dummy_zscore);
fprintf('done\n')
%% EXCLUDE SPECIAL SNPS FROM ANALYSIS
excludevec = false(size(logpvec1,1),1);
fprintf('%i variants are defined across all traits\n', sum(isfinite(logpvec1+sum(logpmat2,2))));
if ~isnan(options.mafthresh)
fprintf('Exclude %i variants due to MAF (minor allele frequency) below %.3f or undefined\n', ...
sum(isnan(mafvec) | (mafvec <= options.mafthresh)), options.mafthresh)
excludevec = excludevec | isnan(mafvec) | (mafvec <= options.mafthresh);
end
for j = 1:length(options.exclude_from_fit)
% Should only exclude for fdr calculations, not discovery
excludevec(options.exclude_from_fit{j}(1):options.exclude_from_fit{j}(2)) = true;
exclude_from = options.exclude_from_fit{j}(1);
exclude_to = options.exclude_from_fit{j}(2);
exclude_chr = unique(chrnumvec(exclude_from:exclude_to));
fprintf('Exclude %i SNPs on chromosome %s from %i to %i KB\n', ...
exclude_to - exclude_from + 1, ...
mat2str(exclude_chr), posvec(exclude_from)/1000, posvec(exclude_to)/1000)
end
excludevec_discovery = false(size(excludevec));
if options.exclude_from_fit_and_discovery
fprintf('Excluded SNPs will be excluded from both fit and discovery')
excludevec_discovery = excludevec;
end
if options.exclude_ambiguous_snps
excludevec = excludevec | is_ambiguous;
excludevec_discovery = excludevec_discovery | is_ambiguous;
fprintf('%i Ambigous SNPs will be excluded from both fit and discovery\n', sum(is_ambiguous))
end
if any(excludevec_discovery)
logpmat2(excludevec_discovery, :) = NaN;
logpvec1(excludevec_discovery) = NaN;
zmat2(excludevec_discovery, :) = NaN;
zvec1(excludevec_discovery) = NaN;
end
if ~any(excludevec), fprintf('No variants were excluded'); end
if any(excludevec), fprintf('%i variants left after exclusion described above\n', ...
sum(~excludevec & isfinite(logpvec1+sum(logpmat2,2))));
end
%% RANDOM PRUNING INDICES
if options.reset_pruneidx || ~exist('pruneidx', 'var'), pruneidx = []; end
if options.randprune && isempty(pruneidx)
% AMD: make sure we only consider variants with defined values when pruning
defvec = ~excludevec & isfinite(logpvec1+sum(logpmat2,2));
pruneidx = random_prune_idx_amd_fb(options.randprune_n, LDmat, defvec, options.randprune_repeats);
end
if options.randprune && (size(pruneidx, 2) ~= options.randprune_n)
error('Invalid pruneidx; try reset_pruneidx=false.');
end
%% GENOMIC CONTROL
if(options.perform_gc)
fprintf('\nPerforming genomic correction... \n')
fprintf('Use %i control variants to calculate lambda GC (genomic correction factor)\n', ...
sum(ivec0 & ~excludevec));
if options.randprune_gc && options.randprune
pruneidx_gc = pruneidx;
fprintf('Random pruning is taken into account in lambdaGC calculation. \n')
else
pruneidx_gc = [];
end
[logpvec1, sig0] = GCcorrect_logpvec(logpvec1, ivec0, options.use_standard_gc, pruneidx_gc);
fprintf('Genomic Control for %s: lambda = %.3f\n', traitname1, median(sig0));
zvec1 = -norminv(10.^-logpvec1/2).*sign(zvec1);
for i=1:length(traitfiles)
[logpmat2(:,i), sig0] = GCcorrect_logpvec(logpmat2(:,i), ivec0, options.use_standard_gc, ...
pruneidx_gc);
fprintf('Genomic Control for %s: lambda = %.3f\n', traitnames{i}, median(sig0));
zmat2(:,i) = -norminv(10.^-logpmat2(:,i)/2).*sign(zmat2(:,i));
end
fprintf('done\n')
else
fprintf('Skipping genomic correction. \n')
end
%% CORRECT FOR SAMPLE OVERLAP
% Decorrelation in Statistics: The Mahalanobis Transformation
% Added material to Data Compression: The Complete Reference
% http://www.davidsalomon.name/DC2advertis/DeCorr.pdf
% Calculation of 'FDRvec0' and 'save_to_csv' should use the original (non-decorrelated) stats
% why? and does this comment actually belong here? -NoF
logpvec1_orig = logpvec1; zvec1_orig = zvec1;
logpmat2_orig = logpmat2; zmat2_orig = zmat2;
if options.correct_for_sample_overlap
[logpvec1, logpmat2] = correct_sample_overlap(logpvec1, logpmat2, ivec0, traitname1, traitnames);
zvec1 = -norminv( 0.5 * (10 .^(-logpvec1) )) .* sign(zvec1);
zmat2 = -norminv( 0.5 * (10 .^(-logpmat2) )) .* sign(zmat2);
else
check_sample_overlap(logpvec1, logpmat2, ivec0, traitname1, traitnames);
end
%% FISHER COMBINED STATISTICS
fprintf('Calculating Fisher combined stats... ')
if (options.fishercomb)
if options.randprune_gc && options.randprune
pruneidx_fc = pruneidx;
elseif options.randprune
pruneidx_fc = [];
end
% AMD: Should not exclude exclude_from_fit variants here; should distinguish from exclude_from_plot!
% flp = fisher_comStats(logpvec1, logpmat2, ivec0, options.use_standard_gc, pruneidx_fc, excludevec);
flp = fisher_comStats(logpvec1, logpmat2, ivec0, options.use_standard_gc, pruneidx_fc);
else
flp = repmat(logpvec1, [1, size(logpmat2,2)]);
end
for iteri=1:size(logpmat2,2)
exclude = isnan(logpmat2(:,iteri));
flp( exclude, iteri) = NaN;
end
fprintf('done\n')
%% PLEIOTROPY ANALYSIS
fprintf('Running pleiotropy analyses... ')
[fdrmat, ~, ~, fdrmat12, fdrmat21, lookup12, lookup21] = ...
pleioFDR_amd(logpvec1, logpmat2, options, LDmat, excludevec, pruneidx);
% the unconditioned fdr is calculated from unpruned variants
[~ , ~, fdrvec0] = cond_FDR_amd(logpvec1_orig, logpmat2(:,1), options, [], []);
fprintf('done\n')
fprintf('mean (std) variants per random pruning iteration = %.2f (%.2f)\n', mean(sum(pruneidx)), ...
std(sum(pruneidx)))
%% PLOT QQ AND ENRICHMENT
fprintf(1,'PLOT QQ / Enrichment... ')
[h_qq, mat_qq] = plot_qq_amd(logpvec1, logpmat2, traitname1, traitnames, options, LDmat, pruneidx, excludevec);
[h_qq_inv, mat_qq_inv] = plot_qq_amd(logpvec1, logpmat2, traitname1, traitnames, options, LDmat, pruneidx, excludevec, true);
[h_tdr, mat_tdr] = plot_qq_amd(logpvec1, logpmat2, traitname1, traitnames, options, LDmat, pruneidx, excludevec, false, false, true);
[h_tdr_inv, mat_tdr_inv] = plot_qq_amd(logpvec1, logpmat2, traitname1, traitnames, options, LDmat, pruneidx, excludevec, true, false, true);
[h_enrich, mat_enrich] = plot_enrichment_amd(logpvec1,logpmat2, traitname1, traitnames, options, LDmat, pruneidx, excludevec);
[h_enrich_inv, mat_enrich_inv] = plot_enrichment_amd(logpvec1,logpmat2, traitname1, traitnames, options, LDmat, pruneidx, excludevec, true);
qq_t1breaks = options.t1breaks;
save(fullfile(options.outputdir, 'result.mat'), '-v6', 'fdrmat', 'logpvec1', 'logpmat2', 'zvec1', 'zmat2', 'excludevec', 'mafvec', 'mat_qq', 'mat_qq_inv', 'mat_enrich', 'mat_enrich_inv', 'qq_t1breaks');
mkdir(outputdir)
filetypes = {'fig','png','svg'};
for j = 1:length(filetypes)
save_figure(h_qq, filetypes{j}, sprintf('%s/%s_vs_%sqq', outputdir, ...
traitname1, sprintf('%s_',traitnames{:})));
save_figure(h_qq_inv, filetypes{j}, sprintf('%s/%svs_%s_qq', outputdir, ...
sprintf('%s_',traitnames{:}), traitname1));
save_figure(h_tdr, filetypes{j}, sprintf('%s/%s_vs_%stdr', outputdir, ...
traitname1, sprintf('%s_',traitnames{:})));
save_figure(h_tdr_inv, filetypes{j}, sprintf('%s/%svs_%s_tdr', outputdir, ...
sprintf('%s_',traitnames{:}), traitname1));
save_figure(h_enrich, filetypes{j}, sprintf('%s/%s_vs_%senrich', outputdir, ...
traitname1, sprintf('%s_',traitnames{:})));
save_figure(h_enrich_inv, filetypes{j}, sprintf('%s/%svs_%s_enrich', outputdir, ...
sprintf('%s_',traitnames{:}), traitname1));
end
fprintf('done\n')
%% PLOT LOOKUP
% AMD: how does plot_lookup use logpvec1 & logpmat2 vs. lookup12 and lookup21?
h_lookup = plot_lookup(logpvec1, logpmat2, traitname1, traitnames, options, LDmat, pruneidx, ...
excludevec, lookup12, lookup21);
filetypes = {'fig','png','svg'};
for j = 1:length(filetypes)
save_figure(h_lookup, filetypes{j}, sprintf('%s/%svs_%s_lookup', outputdir, ...
sprintf('%s_',traitnames{:}), traitname1));
end
%% SAVE FDR-TABLES AND PLOT MANHATTAN AT A SPECIFIC THRESHOLD
if ~exist('snpidlist', 'var'), snpidlist = cell(size(chrnumvec)); snpidlist(:) = {''}; end;
genenamelist = cell(size(chrnumvec)); genenamelist(:) = {''};
if ~exist('A1vec', 'var'), A1vec = cell(size(chrnumvec)); A1vec(:) = {''}; end;
if ~exist('A2vec', 'var'), A2vec = cell(size(chrnumvec)); A2vec(:) = {''}; end;
% FIND LOCI AT GIVEN THRESHOLD
[imat, imat2, logfdrmat] = ind_loci_idx(fdrmat, flp, LDmat, mafvec, options);
[locusnumvec, locusnum] = locusnumber(imat, LDmat, mafvec, options);
results = struct('fdrmat', fdrmat, 'imat', imat, 'imat2', imat2, ...
'logfdrmat', logfdrmat, 'locusnumvec', locusnumvec, 'logpmat2', logpmat2_orig, ...
'logpvec', logpvec1_orig, 'zvec', zvec1_orig, 'zmat2', zmat2_orig, 'fdrvec0', fdrvec0, ...
'snpidlist', {snpidlist}, 'genenamelist', {genenamelist}, ...
'chrnumvec', chrnumvec, 'posvec', posvec, 'A1vec', {A1vec}, 'A2vec', {A2vec}, 'flp', flp, ...
'pruneidx', pruneidx);
if (~is_octave())
disp(results)
end
% SAVE TABLES
fprintf('Saving .csv... ')
save_to_csv(results, options, traitname1, traitnames, outputdir, false); % don't prune CSV
save_to_csv(results, options, traitname1, traitnames, outputdir, true); % prune CSV
fprintf('done\n')
% PLOT MANHATTAN
if options.manh_plot
fprintf('Creating Manhattan plots... ')
h_Manhattan = plot_Manhattan(results, traitname1, traitnames, chrnumvec, options);
switch options.stattype
case 'condfdr'
filetypes = {'fig','png'};
case 'conjfdr'
filetypes = {'fig','png','svg'};
end
for j = 1:length(filetypes)
if strcmp(filetypes{j}, 'png')
save_figure(h_Manhattan, filetypes{j}, sprintf('%s/%s%s_%s_%g_manhattan.png', outputdir, ...
traitname1, sprintf('_%s',traitnames{:}), options.stattype, options.fdrthresh));
else
save_figure(h_Manhattan, filetypes{j}, sprintf('%s/%s%s_%s_%g_manhattan', outputdir, ...
traitname1, sprintf('_%s',traitnames{:}), options.stattype, options.fdrthresh));
end
end
fprintf('done\n')
end
%% CLOSE DIARY
diary off