-
Notifications
You must be signed in to change notification settings - Fork 1
/
oeproteome_enrichment_analysis.jl
267 lines (235 loc) · 13.1 KB
/
oeproteome_enrichment_analysis.jl
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
proj_info = (id = "cov2",
data_ver = "20200527",
fit_ver = "20200608",
oesc_ver = "20200609",
modelobj = "protgroup",
msfolder = "spectronaut_oeproteome_20200527")
using Pkg
Pkg.activate(@__DIR__)
using Revise
using RData, DataFrames
using StatsBase
@info "Project '$(proj_info.id)' dataset $(proj_info.msfolder) version=$(proj_info.data_ver)"
const misc_scripts_path = joinpath(base_scripts_path, "misc_jl");
const analysis_path = joinpath(base_analysis_path, proj_info.id)
const data_path = joinpath(analysis_path, "data")
const results_path = joinpath(analysis_path, "results")
const scratch_path = joinpath(analysis_path, "scratch")
const plots_path = joinpath(analysis_path, "plots")
Revise.includet(joinpath(misc_scripts_path, "frame_utils.jl"));
Revise.includet(joinpath(misc_scripts_path, "msglm_utils.jl"));
objid_col = Symbol(string(proj_info.modelobj, "_id"));
input_rdata = load(joinpath(scratch_path, "$(proj_info.id)_msglm_data_$(proj_info.msfolder)_$(proj_info.fit_ver).RData"), convert=true)
full_rdata = load(joinpath(scratch_path, "$(proj_info.id)_msdata_full_$(proj_info.msfolder)_$(proj_info.data_ver).RData"), convert=true)
fit_rdata = load(joinpath(scratch_path, "$(proj_info.id)_msglm_fit_$(proj_info.msfolder)_$(proj_info.fit_ver).RData"), convert=true)
effects_df = copy(input_rdata["effects.df"]);
contrasts_df = unique!(select!(copy(input_rdata["contrastXmetacondition.df"]), [:contrast, :contrast_type]));
objects_df = copy(input_rdata["msdata"][string(proj_info.modelobj, "s")]) |> MSGLMUtils.fix_object_id!;
objects_df.is_used = objects_df.q_value .<= 0.001
protacs_df = copy(full_rdata["msdata_full"]["proteins"]);
obj2protac_df = select!(
filter(r -> r.is_majority, full_rdata["msdata_full"][string("protein2", proj_info.modelobj)]),
[objid_col, :protein_ac]) |> unique! |> MSGLMUtils.fix_object_id!; # is_majority?
#obj2protac_df = obj2protac_df[obj2protac_df.is_majority, [:object_id, :protein_ac]];
obj_effects_df = filter(r -> r.var == "obj_effect", fit_rdata["object_effects.df"]);
obj_allcontrasts_df = copy(fit_rdata["object_contrasts.df"])
for df in [obj_effects_df, obj_allcontrasts_df]
df |> MSGLMUtils.fix_quantile_columns! |> MSGLMUtils.fix_object_id!
end
sel_std_type = "median"
obj_batchcontrasts_df = filter(r -> r.std_type == sel_std_type &&
occursin(r"_vs_B\d+_others$", r.contrast),
obj_allcontrasts_df)
# use control contrasts by default
obj_contrasts_df = filter(r -> r.std_type == sel_std_type &&
occursin(r"_vs_controls$", r.contrast),
obj_allcontrasts_df)
obj_contrasts_df = innerjoin(obj_contrasts_df,
rename!(select(obj_batchcontrasts_df, [:bait_full_id, :contrast, :std_type, :object_id, :median_log2, :p_value]),
:contrast => :contrast_batch, :median_log2 => :median_log2_batch, :p_value => :p_value_batch),
on = [:bait_full_id, :std_type, :object_id])
obj_contrasts_df = semijoin(obj_contrasts_df, filter(r -> r.is_used, objects_df), on=:object_id)
Revise.includet(joinpath(misc_scripts_path, "gmt_reader.jl"));
Revise.includet(joinpath(misc_scripts_path, "optcover_utils.jl"));
Revise.includet(joinpath(misc_scripts_path, "omics_collections.jl"));
@info "Loading Human annotations..."
# human mappings from http://download.baderlab.org/EM_Genesets/December_01_2018/Human/UniProt/
genesets_df, genesets_coll = GMT.read(String,
joinpath(party3rd_data_path, "Human_GO_AllPathways_with_GO_iea_April_01_2019_UniProt.gmt"),
id_col = :term_id, src_col = :term_src);
pcomplexes_df, pcomplex_iactors_df, pcomplex_iactor2ac_df =
OmicsCollections.ppicollection(joinpath(party3rd_data_path, "complexes_20191217.RData"), seqdb=:uniprot);
pcomplexes_df[!, :coll_id] .= "protein_complexes";
# make complexes collections, keep complexes with at least 2 participants
pcomplex_coll = FrameUtils.frame2collection(join(pcomplex_iactors_df, pcomplex_iactor2ac_df,
on=[:file, :entry_index, :interaction_id, :interactor_id], kind=:inner),
set_col=:complex_id, obj_col=:protein_ac, min_size=2)
protac_sets = merge!(genesets_coll, pcomplex_coll)
terms_df = vcat(rename(genesets_df[!, [:term_src, :term_id, :name, :descr]],
:term_src => :coll_id, :name=>:term_name, :descr=>:term_descr),
#rename(goterm_info_df[[:id, :name, :def]], Dict(:onto => :coll_id, :id=>:term_id, :name=>:term_name, :def=>:term_descr)),
rename(pcomplexes_df[!, [:coll_id, :complex_id, :interaction_label, :interaction_name]],
:complex_id=>:term_id, :interaction_label=>:term_name, :interaction_name=>:term_descr));
protac2term_df = FrameUtils.collection2frame(protac_sets, terms_df,
setid_col=:term_id, objid_col=:protein_ac)
# link protein group IDs to annots and create protgroup collections
obj2term_df = select!(join(obj2protac_df, protac2term_df, on = :protein_ac, kind=:inner),
Not([:protein_ac])) |> unique!
protac_colls = FrameUtils.frame2collections(protac2term_df, obj_col=:protein_ac,
set_col=:term_id, coll_col=:coll_id)
obj_colls = FrameUtils.frame2collections(obj2term_df, obj_col=objid_col,
set_col=:term_id, coll_col=:coll_id)
@info "Preparing hit sets"
ObjectType = eltype(obj2protac_df.object_id)
obj_hit_sets = Dict{Tuple{String, String, String}, Set{ObjectType}}()
obj_contrasts_df.new_change = [
(coalesce(r.p_value, 1.0) <= 0.001 && abs(r.median_log2) >= 0.25) &&
(coalesce(r.p_value_batch, 1.0) <= 0.001 && abs(r.median_log2_batch) >= 0.25) &&
sign(r.median_log2) == sign(r.median_log2_batch) ?
(r.median_log2 > 0 ? "+" : "-") : "."
for r in eachrow(obj_contrasts_df)
]
for hits_df in groupby(filter(r -> r.new_change != ".", obj_contrasts_df), [:std_type, :contrast, :new_change])
obj_hit_sets[(string.(hits_df[1, :std_type], "_std"), hits_df[1, :contrast],
hits_df[1, :new_change])] =
Set(skipmissing(hits_df.object_id))
end
# only relevant ones
obj_hit_selsets = filter(kv ->
(kv[1][1] == sel_std_type*"_std") &&
occursin("_vs_controls", kv[1][2]),
obj_hit_sets);
obj_hit_selsets
ENV["MKL_NUM_THREADS"] = 1
@info "Preparing mosaics..."
observed_protacs = Set(semijoin(obj2protac_df, obj_contrasts_df, on=:object_id).protein_ac) # all annotation ACs observed in the data
obj_mosaics = OptCoverUtils.collections2mosaics(obj_colls, protac_colls, observed_protacs,
setXset_frac_extra_elms=0.05,
verbose=true);
obj_hit_mosaics = Dict(begin
@info "Masking $mosaic_name dataset by hits..."
mosaic_name => OptCoverUtils.automask(mosaic, obj_hit_selsets,
max_sets=2000, min_nmasked=2, max_setsize=200)
end for (mosaic_name, mosaic) in pairs(obj_mosaics));
using OptEnrichedSetCover
cover_params = CoverParams(setXset_factor=0.5,
uncovered_factor=0.0, covered_factor=0.0)#, covered_factor=0.002)
obj_hit_covers = Dict(begin
@info "Covering $mosaic_name by bits..."
mosaic_name => collect(masked_mosaic, cover_params,
CoverEnumerationParams(max_set_score=0.0, max_covers=1),
MultiobjOptimizerParams(ϵ=[0.05, 0.5], MaxSteps=2_000_000, WeightDigits=2,
NWorkers=Threads.nthreads()-1, MaxRestarts=200),
true)
end for (mosaic_name, masked_mosaic) in pairs(obj_hit_mosaics))
using JLD2
@info "Saving data and analysis results"
hit_covers_filename = joinpath(scratch_path, "$(proj_info.id)_contrasts_$(proj_info.oesc_ver)_covers.jld2")
@save(hit_covers_filename,
proj_info, protac_colls, obj_colls, obj_mosaics,
obj2term_df, terms_df,
objects_df, obj_contrasts_df,
obj_hit_sets, obj_hit_selsets, obj_hit_mosaics,
cover_params, obj_hit_covers)
if !@isdefined(obj_hit_covers)
using JLD2, CSV, DataFrames, OptEnrichedSetCover
@load(hit_covers_filename,
proj_info, protac_colls, obj_colls, obj_mosaics,
obj2term_df, terms_df,
objects_df, obj_contrasts_df,
obj_hit_sets, obj_hit_selsets, obj_hit_mosaics,
cover_params, obj_hit_covers)
end
Revise.includet(joinpath(misc_scripts_path, "optcover_utils.jl"));
@info "Preparing protgroup↦gene_name map..."
obj_id2name = Dict(r.object_id => r[Symbol(proj_info.modelobj, "_label")]
for r in eachrow(objects_df))
obj_hit_covers_df = OptCoverUtils.covers_report(
obj_hit_covers, obj_hit_sets, obj_colls, obj_mosaics, obj_id2name,
terms_df,
maskid_col=[:std_type, :contrast, :change],
maskedset_col_prefix="hit")
obj_hit_covers_signif_df = combine(groupby(filter(r -> !occursin("?", r.contrast), obj_hit_covers_df), :term_collection)) do coll_df
@info "Processing $(coll_df.term_collection[1])..."
return select!(OptCoverUtils.filter_multicover(coll_df, set_cols=[:contrast, :change],
max_term_pvalue=5E-5, max_set_pvalue=1E-3, max_entry_pvalue=1.0),
Not(:term_collection))
end
using CSV
CSV.write(joinpath(analysis_path, "reports", "$(proj_info.id)_$(proj_info.msfolder)_hit_oesc_$(sel_std_type)_std_$(proj_info.oesc_ver).txt"),
obj_hit_covers_df[obj_hit_covers_df.nmasked .> 0, :],
missingstring="", delim='\t');
CSV.write(joinpath(analysis_path, "reports", "$(proj_info.id)_$(proj_info.msfolder)_hit_oesc_$(sel_std_type)_std_signif_$(proj_info.oesc_ver).txt"),
obj_hit_covers_signif_df[obj_hit_covers_signif_df.nmasked .> 0, :],
missingstring="", delim='\t');
Revise.includet(joinpath(misc_scripts_path, "frame_utils.jl"))
Revise.includet(joinpath(misc_scripts_path, "optcover_plots.jl"))
Revise.includet(joinpath(misc_scripts_path, "optcover_heatmap.jl"))
using PlotlyJS, TextWrap, ORCA
heatmap_layout_attrs = Dict(
("GO_CC", true) => Dict(:margin_l => 200),
("GO_CC", false) => Dict(:margin_l => 200),
)
#=
for (plot_mosaic, cover_coll) in obj_hit_covers
isempty(cover_coll.results) && continue
@info "Plotting $plot_mosaic Pareto front"
paretofront_plot = OptCoverPlots.plot_paretofront(cover_coll.results[1], plot_unfolded=true)
plot_filename = joinpath(plots_path, "oesc_$(sel_std_type)_std", "paretofront",
"$(proj_info.id)_$(plot_mosaic)_X_treatment_$(sel_std_type)_pareto")
savefig(paretofront_plot.plot, "$plot_filename.svg")
PlotlyJS.savehtml(paretofront_plot, "$plot_filename.html")
end
=#
stylize_contrast(str) = foldl(replace, [
"_vs_controls" => "",#" <span style=\"color: #808080;\">vs</span> ",
r"SARS_CoV2_(\w+)" => s"\1<sub><span style=\"color: orange;\">CoV2</span></sub>",
r"SARS_CoV_(\w+)" => s"\1<sub><span style=\"color: sienna;\">CoV</span></sub>",
r"HCoV_(\w+)" => s"\1<sub><span style=\"color: seagreen;\">HCoV</span></sub>",
#"controls" => "<span style=\"color: #808080;\">controls</span>",
#"others" => "<span style=\"color: #808080;\">others</span>",
],
init = str)
function process_contrast_axis(contrast_df)
contrast_df,
string.(stylize_contrast.(contrast_df.contrast), " ",
OptCoverHeatmap.stylize_change.(contrast_df.change)),
string.(stylize_contrast.(contrast_df.contrast), " ",
OptCoverHeatmap.stylize_change.(contrast_df.change))
end
for term_coll in unique(obj_hit_covers_df.term_collection), signif in (false, true)
@info "Plotting $(signif ? "signif " : "")heatmap for $term_coll..."
layout_attrs = get(heatmap_layout_attrs, (term_coll, signif), Dict())
df = signif ? obj_hit_covers_signif_df : obj_hit_covers_df
for outformat in ["svg", "html"]# ("html", "pdf", "svg")
coll_heatmap = OptCoverHeatmap.oesc_heatmap(df,
Symbol(term_coll), elements_label="proteins",
maskedset_axis_title = "Comparison",
maskedset_cols = [:contrast, :change, :nhit],
process_maskedset_axis=process_contrast_axis,
process_term_axis=OptCoverHeatmap.process_term_axis,
margin_l=get(layout_attrs, :margin_l, 400),
margin_b=get(layout_attrs, :margin_b, 60),
colorscale = "Hot", reversescale=false,
plot_bgcolor="#FFF", gridcolor="#DDD",#outformat in ["svg", "pdf"] ? "#000" : "#BBB",
cell_width=25, cell_height=14, gridwidth=1,
transpose=false)
(coll_heatmap === nothing) && continue
for (k, v) in [#:width=>800, :height=>400,
:margin_r=>100, :margin_t=>20,
:yaxis_tickfont_size=>12, :xaxis_tickangle=>45]
coll_heatmap.plot.layout[k] = v
end
plot_fname = joinpath(plots_path, "$(proj_info.msfolder)_$(proj_info.fit_ver)", "oesc_hits_$(sel_std_type)",
"$(proj_info.id)_$(proj_info.oesc_ver)_$(term_coll)_X_hits$(signif ? "_signif" : "")_heatmap.$(outformat)")
if outformat == "html"
PlotlyJS.savehtml(coll_heatmap, plot_fname, :embed);
else
try
savefig(coll_heatmap.plot, plot_fname);
catch e
@warn e
end
end
end
end