aboutsummaryrefslogtreecommitdiff
path: root/gn3/heatmaps.py
blob: c48a2d3e7d449562f4d1b3737640374740b55a26 (plain)
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
"""
This module will contain functions to be used in computation of the data used to
generate various kinds of heatmaps.
"""

from functools import reduce
from gn3.settings import TMPDIR
import plotly.graph_objects as go
from gn3.random import random_string
from typing import Any, Dict, Sequence
from gn3.computations.slink import slink
from plotly.subplots import make_subplots
from gn3.computations.correlations2 import compute_correlation
from gn3.db.genotypes import (
    build_genotype_file, load_genotype_samples, parse_genotype_file)
from gn3.db.traits import (
    retrieve_trait_data,
    retrieve_trait_info,
    generate_traits_filename)
from gn3.computations.qtlreaper import (
    run_reaper,
    generate_traits_file,
    chromosome_sorter_key_fn,
    parse_reaper_main_results,
    organise_reaper_main_results,
    parse_reaper_permutation_results)

def export_trait_data(
        trait_data: dict, strainlist: Sequence[str], dtype: str = "val",
        var_exists: bool = False, n_exists: bool = False):
    """
    Export data according to `strainlist`. Mostly used in calculating
    correlations.

    DESCRIPTION:
    Migrated from
    https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/base/webqtlTrait.py#L166-L211

    PARAMETERS
    trait: (dict)
      The dictionary of key-value pairs representing a trait
    strainlist: (list)
      A list of strain names
    dtype: (str)
      ... verify what this is ...
    var_exists: (bool)
      A flag indicating existence of variance
    n_exists: (bool)
      A flag indicating existence of ndata
    """
    def __export_all_types(tdata, strain):
        sample_data = []
        if tdata[strain]["value"]:
            sample_data.append(tdata[strain]["value"])
            if var_exists:
                if tdata[strain]["variance"]:
                    sample_data.append(tdata[strain]["variance"])
                else:
                    sample_data.append(None)
            if n_exists:
                if tdata[strain]["ndata"]:
                    sample_data.append(tdata[strain]["ndata"])
                else:
                    sample_data.append(None)
        else:
            if var_exists and n_exists:
                sample_data += [None, None, None]
            elif var_exists or n_exists:
                sample_data += [None, None]
            else:
                sample_data.append(None)

        return tuple(sample_data)

    def __exporter(accumulator, strain):
        # pylint: disable=[R0911]
        if strain in trait_data["data"]:
            if dtype == "val":
                return accumulator + (trait_data["data"][strain]["value"], )
            if dtype == "var":
                return accumulator + (trait_data["data"][strain]["variance"], )
            if dtype == "N":
                return accumulator + (trait_data["data"][strain]["ndata"], )
            if dtype == "all":
                return accumulator + __export_all_types(trait_data["data"], strain)
            raise KeyError("Type `%s` is incorrect" % dtype)
        if var_exists and n_exists:
            return accumulator + (None, None, None)
        if var_exists or n_exists:
            return accumulator + (None, None)
        return accumulator + (None,)

    return reduce(__exporter, strainlist, tuple())

def trait_display_name(trait: Dict):
    """
    Given a trait, return a name to use to display the trait on a heatmap.

    DESCRIPTION
    Migrated from
    https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/base/webqtlTrait.py#L141-L157
    """
    if trait.get("db", None) and trait.get("trait_name", None):
        if trait["db"]["dataset_type"] == "Temp":
            desc = trait["description"]
            if desc.find("PCA") >= 0:
                return "%s::%s" % (
                    trait["db"]["displayname"],
                    desc[desc.rindex(':')+1:].strip())
            return "%s::%s" % (
                trait["db"]["displayname"],
                desc[:desc.index('entered')].strip())
        prefix = "%s::%s" % (
            trait["db"]["dataset_name"], trait["trait_name"])
        if trait["cellid"]:
            return "%s::%s" % (prefix, trait["cellid"])
        return prefix
    return trait["description"]

def cluster_traits(traits_data_list: Sequence[Dict]):
    """
    Clusters the trait values.

    DESCRIPTION
    Attempts to replicate the clustering of the traits, as done at
    https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/heatmap/Heatmap.py#L138-L162
    """
    def __compute_corr(tdata_i, tdata_j):
        if tdata_i[0] == tdata_j[0]:
            return 0.0
        corr_vals = compute_correlation(tdata_i[1], tdata_j[1])
        corr = corr_vals[0]
        if (1 - corr) < 0:
            return 0.0
        return 1 - corr

    def __cluster(tdata_i):
        return tuple(
            __compute_corr(tdata_i, tdata_j)
            for tdata_j in enumerate(traits_data_list))

    return tuple(__cluster(tdata_i) for tdata_i in enumerate(traits_data_list))

def build_heatmap(traits_names, conn: Any):
    """
    heatmap function

    DESCRIPTION
    This function is an attempt to reproduce the initialisation at
    https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/heatmap/Heatmap.py#L46-L64
    and also the clustering and slink computations at
    https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/heatmap/Heatmap.py#L138-L165
    with the help of the `gn3.computations.heatmap.cluster_traits` function.

    It does not try to actually draw the heatmap image.

    PARAMETERS:
    TODO: Elaborate on the parameters here...
    """
    threshold = 0 # webqtlConfig.PUBLICTHRESH
    traits = [
        retrieve_trait_info(threshold, fullname, conn)
        for fullname in trait_fullnames()]
    traits_data_list = [retrieve_trait_data(t, conn) for t in traits]
    genotype_filename = build_genotype_file(traits[0]["riset"])
    genotype = parse_genotype_file(genotype_filename)
    strains = load_genotype_samples(genotype_filename)
    exported_traits_data_list = [
        export_trait_data(td, strains) for td in traits_data_list]
    slinked = slink(cluster_traits(exported_traits_data_list))
    traits_order = compute_traits_order(slinked)
    ordered_traits_names = [
        traits[idx]["trait_fullname"] for idx in traits_order]
    strains_and_values = retrieve_strains_and_values(
        traits_order, strains, exported_traits_data_list)
    traits_filename = "{}/traits_test_file_{}.txt".format(
        TMPDIR, random_string(10))
    generate_traits_file(
        strains_and_values[0][1],
        [t[2] for t in strains_and_values],
        traits_filename)

    main_output, permutations_output = run_reaper(
        genotype_filename, traits_filename, separate_nperm_output=True)

    qtlresults = parse_reaper_main_results(main_output)
    permudata = parse_reaper_permutation_results(permutations_output)
    organised = organise_reaper_main_results(qtlresults)

    traits_ids = [# sort numerically, but retain the ids as strings
        str(i) for i in sorted({int(row["ID"]) for row in qtlresults})]
    chromosome_names = sorted(
        {row["Chr"] for row in qtlresults}, key = chromosome_sorter_key_fn)
    loci_names = sorted({row["Locus"] for row in qtlresults})
    ordered_traits_names = {
        res_id: trait for res_id, trait in
        zip(traits_ids,
            [traits[idx]["trait_fullname"] for idx in traits_order])}

    return generate_clustered_heatmap(
        process_traits_data_for_heatmap(
            organised, traits_ids, chromosome_names),
        "single_heatmap_{}".format(random_string(10)),
        y_axis=tuple(
            ordered_traits_names[traits_ids[order]]
                for order in traits_order),
        y_label="Traits",
        x_axis=[chromo for chromo in chromosome_names],
        x_label="Chromosomes")

    return {
        "slink_data": slink_data,
        "ordering_data": ordering_data,
        "strainlist": strainlist,
        "genotype_filename": genotype_filename,
        "traits_list": traits_list,
        "traits_data_list": traits_data_list,
        "exported_traits_data_list": exported_traits_data_list,
        "traits_filename": traits_filename
    }

def compute_traits_order(slink_data, neworder: tuple = tuple()):
    """
    Compute the order of the traits for clustering from `slink_data`.

    This function tries to reproduce the creation and update of the `neworder`
    variable in
    https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/heatmap/Heatmap.py#L120
    and in the `web.webqtl.heatmap.Heatmap.draw` function in GN1
    """
    def __order_maker(norder, slnk_dt):
        if isinstance(slnk_dt[0], int) and isinstance(slnk_dt[1], int):
            return norder + (slnk_dt[0], slnk_dt[1])

        if isinstance(slnk_dt[0], int):
            return __order_maker((norder + (slnk_dt[0], )), slnk_dt[1])

        if isinstance(slnk_dt[1], int):
            return __order_maker(norder, slnk_dt[0]) + (slnk_dt[1], )

        return __order_maker(__order_maker(norder, slnk_dt[0]), slnk_dt[1])

    return __order_maker(neworder, slink_data)

def retrieve_strains_and_values(orders, strainlist, traits_data_list):
    """
    Get the strains and their corresponding values from `strainlist` and
    `traits_data_list`.

    This migrates the code in
    https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/heatmap/Heatmap.py#L215-221
    """
    # This feels nasty! There's a lot of mutation of values here, that might
    # indicate something untoward in the design of this function and its
    # dependents  ==>  Review
    strains = []
    values = []
    rets = []
    for order in orders:
        temp_val = traits_data_list[order]
        for i, strain in enumerate(strainlist):
            if temp_val[i] is not None:
                strains.append(strain)
                values.append(temp_val[i])
        rets.append([order, strains[:], values[:]])
        strains = []
        values = []

    return rets

def nearest_marker_finder(genotype):
    """
    Returns a function to be used with `genotype` to compute the nearest marker
    to the trait passed to the returned function.

    https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/heatmap/Heatmap.py#L425-434
    """
    def __compute_distances(chromo, trait):
        loci = chromo.get("loci", None)
        if not loci:
            return None
        return tuple(
            {
                "name": locus["name"],
                "distance": abs(locus["Mb"] - trait["mb"])
            } for locus in loci)

    def __finder(trait):
        _chrs = tuple(
            _chr for _chr in genotype["chromosomes"]
            if str(_chr["name"]) == str(trait["chr"]))
        if len(_chrs) == 0:
            return None
        distances = tuple(
            distance for dists in
            filter(
                lambda x: x is not None,
                (__compute_distances(_chr, trait) for _chr in _chrs))
            for distance in dists)
        nearest = min(distances, key=lambda d: d["distance"])
        return nearest["name"]
    return __finder

def get_nearest_marker(traits_list, genotype):
    """
    Retrieves the nearest marker for each of the traits in the list.

    DESCRIPTION:
    This migrates the code in
    https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/heatmap/Heatmap.py#L419-L438
    """
    if not genotype["Mbmap"]:
        return [None] * len(trait_list)

    marker_finder = nearest_marker_finder(genotype)
    return [marker_finder(trait) for trait in traits_list]

def get_lrs_from_chr(trait, chr_name):
    """
    Retrieve the LRS values for a specific chromosome in the given trait.
    """
    chromosome = trait["chromosomes"].get(chr_name)
    if chromosome:
        return [
            locus["LRS"] for locus in
            sorted(chromosome["loci"], key=lambda loc: loc["Locus"])]
    return [None]

def process_traits_data_for_heatmap(data, trait_names, chromosome_names):
    """
    Process the traits data in a format useful for generating heatmap diagrams.
    """
    hdata = [
        [get_lrs_from_chr(data[trait], chr_name) for trait in trait_names]
        for chr_name in chromosome_names]
    return hdata

def generate_clustered_heatmap(
        data, image_filename_prefix, x_axis = None, x_label: str = "",
        y_axis = None, y_label: str = "", output_dir: str = TMPDIR,
        colorscale = [
            [0.0, '#3B3B3B'], [0.4999999999999999, '#ABABAB'],
            [0.5, '#F5DE11']], [1.0, '#FF0D00']):
    """
    Generate a dendrogram, and heatmaps for each chromosome, and put them all
    into one plot.
    """
    num_cols = len(x_axis)
    fig = make_subplots(
        rows=1,
        cols=num_cols,
        shared_yaxes="rows",
        # horizontal_spacing=(1 / (num_cols - 1)),
        subplot_titles=x_axis
    )
    hms = [go.Heatmap(
        name=chromo,
        y = y_axis,
        z = data_array,
        showscale=False) for chromo, data_array in zip(x_axis, data)]
    for col, hm in enumerate(hms):
        fig.add_trace(hm, row=1, col=(col + 1))

    fig.update_traces(
        showlegend=False,
        colorscale=colorscale,
        selector={"type": "heatmap"})
    fig.update_traces(
        showlegend=True,
        showscale=True,
        selector={"name": x_axis[-1]})
    fig.update_layout(
        coloraxis_colorscale=colorscale
    )
    image_filename = "{}/{}.html".format(output_dir, image_filename_prefix)
    fig.write_html(image_filename)
    return image_filename, fig