| 1 | #! /usr/bin/env python |
|---|
| 2 | |
|---|
| 3 | import sys |
|---|
| 4 | import os |
|---|
| 5 | import time |
|---|
| 6 | import glob |
|---|
| 7 | import multiprocessing |
|---|
| 8 | from operator import itemgetter |
|---|
| 9 | from subprocess import call |
|---|
| 10 | |
|---|
| 11 | from epac.ete2 import Tree, SeqGroup |
|---|
| 12 | from epac.argparse import ArgumentParser,RawDescriptionHelpFormatter |
|---|
| 13 | from epac.config import SativaConfig,EpacConfig |
|---|
| 14 | from epac.raxml_util import RaxmlWrapper, FileUtils |
|---|
| 15 | from epac.json_util import RefJsonParser, RefJsonChecker, EpaJsonParser |
|---|
| 16 | from epac.taxonomy_util import TaxCode, Taxonomy |
|---|
| 17 | from epac.classify_util import TaxTreeHelper,TaxClassifyHelper |
|---|
| 18 | import epa_trainer |
|---|
| 19 | |
|---|
| 20 | class LeaveOneTest: |
|---|
| 21 | def __init__(self, config): |
|---|
| 22 | self.cfg = config |
|---|
| 23 | |
|---|
| 24 | self.mis_fname = self.cfg.out_fname("%NAME%.mis") |
|---|
| 25 | self.premis_fname = self.cfg.out_fname("%NAME%.premis") |
|---|
| 26 | self.misrank_fname = self.cfg.out_fname("%NAME%.misrank") |
|---|
| 27 | self.stats_fname = self.cfg.out_fname("%NAME%.stats") |
|---|
| 28 | |
|---|
| 29 | if os.path.isfile(self.mis_fname): |
|---|
| 30 | print "\nERROR: Output file already exists: %s" % self.mis_fname |
|---|
| 31 | print "Please specify a different job name using -n or remove old output files." |
|---|
| 32 | self.cfg.exit_user_error() |
|---|
| 33 | |
|---|
| 34 | self.tmp_refaln = config.tmp_fname("%NAME%.refaln") |
|---|
| 35 | self.reftree_lbl_fname = config.tmp_fname("%NAME%_lbl.tre") |
|---|
| 36 | self.reftree_tax_fname = config.tmp_fname("%NAME%_tax.tre") |
|---|
| 37 | self.optmod_fname = self.cfg.tmp_fname("%NAME%.opt") |
|---|
| 38 | self.reftree_fname = self.cfg.tmp_fname("ref_%NAME%.tre") |
|---|
| 39 | |
|---|
| 40 | self.mislabels = [] |
|---|
| 41 | self.mislabels_cnt = [] |
|---|
| 42 | self.rank_mislabels = [] |
|---|
| 43 | self.rank_mislabels_cnt = [] |
|---|
| 44 | self.misrank_conf_map = {} |
|---|
| 45 | |
|---|
| 46 | def write_bid_tax_map(self, bid_tax_map, final): |
|---|
| 47 | if self.cfg.debug: |
|---|
| 48 | fname_suffix = "final" if final else "l1out" |
|---|
| 49 | bid_fname = self.cfg.tmp_fname("%NAME%_" + "bid_tax_map_%s.txt" % fname_suffix) |
|---|
| 50 | with open(bid_fname, "w") as outf: |
|---|
| 51 | for bid, bid_rec in bid_tax_map.iteritems(): |
|---|
| 52 | outf.write("%s\t%s\t%d\t%f\n" % (bid, bid_rec[0], bid_rec[1], bid_rec[2])); |
|---|
| 53 | |
|---|
| 54 | def write_assignments(self, assign_map, final): |
|---|
| 55 | if self.cfg.debug: |
|---|
| 56 | fname_suffix = "final" if final else "l1out" |
|---|
| 57 | assign_fname = self.cfg.tmp_fname("%NAME%_" + "taxassign_%s.txt" % fname_suffix) |
|---|
| 58 | with open(assign_fname, "w") as outf: |
|---|
| 59 | for seq_name in assign_map.iterkeys(): |
|---|
| 60 | ranks, lws = assign_map[seq_name] |
|---|
| 61 | outf.write("%s\t%s\t%s\n" % (seq_name, ";".join(ranks), ";".join(["%.3f" % l for l in lws]))) |
|---|
| 62 | |
|---|
| 63 | def load_refjson(self, refjson_fname): |
|---|
| 64 | try: |
|---|
| 65 | self.refjson = RefJsonParser(refjson_fname) |
|---|
| 66 | except ValueError: |
|---|
| 67 | self.cfg.exit_user_error("ERROR: Invalid json file format!") |
|---|
| 68 | |
|---|
| 69 | #validate input json format |
|---|
| 70 | (valid, err) = self.refjson.validate() |
|---|
| 71 | if not valid: |
|---|
| 72 | self.cfg.log.error("ERROR: Parsing reference JSON file failed:\n%s", err) |
|---|
| 73 | self.cfg.exit_user_error() |
|---|
| 74 | |
|---|
| 75 | self.rate = self.refjson.get_rate() |
|---|
| 76 | self.node_height = self.refjson.get_node_height() |
|---|
| 77 | self.origin_taxonomy = self.refjson.get_origin_taxonomy() |
|---|
| 78 | self.tax_tree = self.refjson.get_tax_tree() |
|---|
| 79 | self.cfg.compress_patterns = self.refjson.get_pattern_compression() |
|---|
| 80 | |
|---|
| 81 | self.bid_taxonomy_map = self.refjson.get_branch_tax_map() |
|---|
| 82 | if not self.bid_taxonomy_map: |
|---|
| 83 | # old file format (before 1.6), need to rebuild this map from scratch |
|---|
| 84 | th = TaxTreeHelper(self.cfg, self.origin_taxonomy) |
|---|
| 85 | th.set_mf_rooted_tree(self.tax_tree) |
|---|
| 86 | th.set_bf_unrooted_tree(self.refjson.get_reftree()) |
|---|
| 87 | self.bid_taxonomy_map = th.get_bid_taxonomy_map() |
|---|
| 88 | |
|---|
| 89 | self.write_bid_tax_map(self.bid_taxonomy_map, final=False) |
|---|
| 90 | |
|---|
| 91 | reftree_str = self.refjson.get_raxml_readable_tree() |
|---|
| 92 | self.reftree = Tree(reftree_str) |
|---|
| 93 | self.reftree_size = len(self.reftree.get_leaves()) |
|---|
| 94 | |
|---|
| 95 | # IMPORTANT: set EPA heuristic rate based on tree size! |
|---|
| 96 | self.cfg.resolve_auto_settings(self.reftree_size) |
|---|
| 97 | # If we're loading the pre-optimized model, we MUST set the same rate het. mode as in the ref file |
|---|
| 98 | if self.cfg.epa_load_optmod: |
|---|
| 99 | self.cfg.raxml_model = self.refjson.get_ratehet_model() |
|---|
| 100 | |
|---|
| 101 | self.classify_helper = TaxClassifyHelper(self.cfg, self.bid_taxonomy_map, self.rate, self.node_height) |
|---|
| 102 | self.taxtree_helper = TaxTreeHelper(self.cfg, self.origin_taxonomy, self.tax_tree) |
|---|
| 103 | |
|---|
| 104 | tax_code_name = self.refjson.get_taxcode() |
|---|
| 105 | self.tax_code = TaxCode(tax_code_name) |
|---|
| 106 | |
|---|
| 107 | self.taxonomy = Taxonomy(prefix=EpacConfig.REF_SEQ_PREFIX, tax_map=self.origin_taxonomy) |
|---|
| 108 | self.tax_common_ranks = self.taxonomy.get_common_ranks() |
|---|
| 109 | # print "Common ranks: ", self.tax_common_ranks |
|---|
| 110 | |
|---|
| 111 | self.mislabels_cnt = [0] * TaxCode.UNI_TAX_LEVELS |
|---|
| 112 | self.rank_mislabels_cnt = [0] * TaxCode.UNI_TAX_LEVELS |
|---|
| 113 | |
|---|
| 114 | def run_epa_trainer(self): |
|---|
| 115 | epa_trainer.run_trainer(self.cfg) |
|---|
| 116 | |
|---|
| 117 | if not os.path.isfile(self.cfg.refjson_fname): |
|---|
| 118 | self.cfg.log.error("\nBuilding reference tree failed, see error messages above.") |
|---|
| 119 | self.cfg.exit_fatal_error() |
|---|
| 120 | |
|---|
| 121 | def classify_seq(self, placement): |
|---|
| 122 | edges = placement["p"] |
|---|
| 123 | if len(edges) > 0: |
|---|
| 124 | return self.classify_helper.classify_seq(edges) |
|---|
| 125 | else: |
|---|
| 126 | print "ERROR: no placements! something is definitely wrong!" |
|---|
| 127 | |
|---|
| 128 | def check_seq_tax_labels(self, seq_name, orig_ranks, ranks, lws): |
|---|
| 129 | mis_rec = None |
|---|
| 130 | |
|---|
| 131 | num_common_ranks = len(self.tax_common_ranks) |
|---|
| 132 | orig_rank_level = Taxonomy.lowest_assigned_rank_level(orig_ranks) |
|---|
| 133 | new_rank_level = Taxonomy.lowest_assigned_rank_level(ranks) |
|---|
| 134 | #if new_rank_level < 0 or (new_rank_level < num_common_ranks and orig_rank_level >= num_common_ranks): |
|---|
| 135 | # if new_rank_level < 0: |
|---|
| 136 | if len(ranks) == 0: |
|---|
| 137 | mis_rec = {} |
|---|
| 138 | mis_rec['name'] = seq_name |
|---|
| 139 | mis_rec['orig_level'] = -1 |
|---|
| 140 | mis_rec['real_level'] = 0 |
|---|
| 141 | mis_rec['level_name'] = "[NotIngroup]" |
|---|
| 142 | mis_rec['inv_level'] = -1 * mis_rec['real_level'] # just for sorting |
|---|
| 143 | mis_rec['orig_ranks'] = orig_ranks |
|---|
| 144 | mis_rec['ranks'] = [] |
|---|
| 145 | mis_rec['lws'] = [1.0] |
|---|
| 146 | mis_rec['conf'] = mis_rec['lws'][0] |
|---|
| 147 | else: |
|---|
| 148 | mislabel_lvl = -1 |
|---|
| 149 | min_len = min(len(orig_ranks),len(ranks)) |
|---|
| 150 | for rank_lvl in range(min_len): |
|---|
| 151 | if ranks[rank_lvl] != Taxonomy.EMPTY_RANK and ranks[rank_lvl] != orig_ranks[rank_lvl]: |
|---|
| 152 | mislabel_lvl = rank_lvl |
|---|
| 153 | break |
|---|
| 154 | |
|---|
| 155 | if mislabel_lvl >= 0: |
|---|
| 156 | real_lvl = self.tax_code.guess_rank_level(orig_ranks, mislabel_lvl) |
|---|
| 157 | mis_rec = {} |
|---|
| 158 | mis_rec['name'] = seq_name |
|---|
| 159 | mis_rec['orig_level'] = mislabel_lvl |
|---|
| 160 | mis_rec['real_level'] = real_lvl |
|---|
| 161 | mis_rec['level_name'] = self.tax_code.rank_level_name(real_lvl)[0] |
|---|
| 162 | mis_rec['inv_level'] = -1 * mis_rec['real_level'] # just for sorting |
|---|
| 163 | mis_rec['orig_ranks'] = orig_ranks |
|---|
| 164 | mis_rec['ranks'] = ranks |
|---|
| 165 | mis_rec['lws'] = lws |
|---|
| 166 | mis_rec['conf'] = lws[mislabel_lvl] |
|---|
| 167 | |
|---|
| 168 | if mis_rec: |
|---|
| 169 | self.mislabels.append(mis_rec) |
|---|
| 170 | |
|---|
| 171 | return mis_rec |
|---|
| 172 | |
|---|
| 173 | def filter_mislabels(self): |
|---|
| 174 | filtered_mis = [] |
|---|
| 175 | for i in range(len(self.mislabels)): |
|---|
| 176 | if self.mislabels[i]['conf'] >= self.cfg.conf_cutoff: |
|---|
| 177 | filtered_mis.append(self.mislabels[i]) |
|---|
| 178 | |
|---|
| 179 | self.mislabels = filtered_mis |
|---|
| 180 | |
|---|
| 181 | def check_rank_tax_labels(self, rank_name, orig_ranks, ranks, lws): |
|---|
| 182 | mislabel_lvl = -1 |
|---|
| 183 | min_len = min(len(orig_ranks),len(ranks)) |
|---|
| 184 | for rank_lvl in range(min_len): |
|---|
| 185 | if ranks[rank_lvl] != Taxonomy.EMPTY_RANK and ranks[rank_lvl] != orig_ranks[rank_lvl]: |
|---|
| 186 | mislabel_lvl = rank_lvl |
|---|
| 187 | break |
|---|
| 188 | |
|---|
| 189 | if mislabel_lvl >= 0: |
|---|
| 190 | real_lvl = self.tax_code.guess_rank_level(orig_ranks, mislabel_lvl) |
|---|
| 191 | mis_rec = {} |
|---|
| 192 | mis_rec['name'] = rank_name |
|---|
| 193 | mis_rec['orig_level'] = mislabel_lvl |
|---|
| 194 | mis_rec['real_level'] = real_lvl |
|---|
| 195 | mis_rec['level_name'] = self.tax_code.rank_level_name(real_lvl)[0] |
|---|
| 196 | mis_rec['inv_level'] = -1 * real_lvl # just for sorting |
|---|
| 197 | mis_rec['orig_ranks'] = orig_ranks |
|---|
| 198 | mis_rec['ranks'] = ranks |
|---|
| 199 | mis_rec['lws'] = lws |
|---|
| 200 | mis_rec['conf'] = lws[mislabel_lvl] |
|---|
| 201 | self.rank_mislabels.append(mis_rec) |
|---|
| 202 | |
|---|
| 203 | return mis_rec |
|---|
| 204 | else: |
|---|
| 205 | return None |
|---|
| 206 | |
|---|
| 207 | def mis_rec_to_string_old(self, mis_rec): |
|---|
| 208 | lvl = mis_rec['orig_level'] |
|---|
| 209 | output = mis_rec['name'] + "\t" |
|---|
| 210 | output += "%s\t%s\t%s\t%.3f\n" % (mis_rec['level_name'], |
|---|
| 211 | mis_rec['orig_ranks'][lvl], mis_rec['ranks'][lvl], mis_rec['lws'][lvl]) |
|---|
| 212 | output += ";".join(mis_rec['orig_ranks']) + "\n" |
|---|
| 213 | output += ";".join(mis_rec['ranks']) + "\n" |
|---|
| 214 | output += "\t".join(["%.3f" % conf for conf in mis_rec['lws']]) + "\n" |
|---|
| 215 | return output |
|---|
| 216 | |
|---|
| 217 | def mis_rec_to_string(self, mis_rec): |
|---|
| 218 | lvl = mis_rec['orig_level'] |
|---|
| 219 | uncorr_name = EpacConfig.strip_ref_prefix(self.refjson.get_uncorr_seqid(mis_rec['name'])) |
|---|
| 220 | uncorr_orig_ranks = self.refjson.get_uncorr_ranks(mis_rec['orig_ranks']) |
|---|
| 221 | uncorr_ranks = self.refjson.get_uncorr_ranks(mis_rec['ranks']) |
|---|
| 222 | output = uncorr_name + "\t" |
|---|
| 223 | |
|---|
| 224 | if lvl >= 0: |
|---|
| 225 | output += "%s\t%s\t%s\t%.3f\t" % (mis_rec['level_name'], |
|---|
| 226 | uncorr_orig_ranks[lvl], uncorr_ranks[lvl], mis_rec['lws'][lvl]) |
|---|
| 227 | else: |
|---|
| 228 | output += "%s\t%s\t%s\t%.3f\t" % (mis_rec['level_name'], |
|---|
| 229 | "NA", "NA", mis_rec['lws'][0]) |
|---|
| 230 | |
|---|
| 231 | output += Taxonomy.lineage_str(uncorr_orig_ranks) + "\t" |
|---|
| 232 | output += Taxonomy.lineage_str(uncorr_ranks) + "\t" |
|---|
| 233 | output += ";".join(["%.3f" % conf for conf in mis_rec['lws']]) |
|---|
| 234 | if 'rank_conf' in mis_rec: |
|---|
| 235 | output += "\t%.3f" % mis_rec['rank_conf'] |
|---|
| 236 | return output |
|---|
| 237 | |
|---|
| 238 | def sort_mislabels(self): |
|---|
| 239 | self.mislabels = sorted(self.mislabels, key=itemgetter('inv_level', 'conf', 'name'), reverse=True) |
|---|
| 240 | for mis_rec in self.mislabels: |
|---|
| 241 | real_lvl = mis_rec["real_level"] |
|---|
| 242 | self.mislabels_cnt[real_lvl] += 1 |
|---|
| 243 | |
|---|
| 244 | if self.cfg.ranktest: |
|---|
| 245 | self.rank_mislabels = sorted(self.rank_mislabels, key=itemgetter('inv_level', 'conf', 'name'), reverse=True) |
|---|
| 246 | for mis_rec in self.rank_mislabels: |
|---|
| 247 | real_lvl = mis_rec["real_level"] |
|---|
| 248 | self.rank_mislabels_cnt[real_lvl] += 1 |
|---|
| 249 | |
|---|
| 250 | def write_stats(self, toFile=False): |
|---|
| 251 | self.cfg.log.info("Mislabeled sequences by rank:") |
|---|
| 252 | seq_sum = 0 |
|---|
| 253 | rank_sum = 0 |
|---|
| 254 | stats = [] |
|---|
| 255 | for i in range(len(self.mislabels_cnt)): |
|---|
| 256 | if i > 0: |
|---|
| 257 | rname = self.tax_code.rank_level_name(i)[0].ljust(12) |
|---|
| 258 | else: |
|---|
| 259 | rname = "[NotIngroup]" |
|---|
| 260 | if self.mislabels_cnt[i] > 0: |
|---|
| 261 | seq_sum += self.mislabels_cnt[i] |
|---|
| 262 | # output = "%s:\t%d" % (rname, seq_sum) |
|---|
| 263 | output = "%s:\t%d" % (rname, self.mislabels_cnt[i]) |
|---|
| 264 | if self.cfg.ranktest: |
|---|
| 265 | rank_sum += self.rank_mislabels_cnt[i] |
|---|
| 266 | output += "\t%d" % rank_sum |
|---|
| 267 | self.cfg.log.info(output) |
|---|
| 268 | stats.append(output) |
|---|
| 269 | |
|---|
| 270 | if toFile: |
|---|
| 271 | with open(self.stats_fname, "w") as fo_stat: |
|---|
| 272 | for line in stats: |
|---|
| 273 | fo_stat.write(line + "\n") |
|---|
| 274 | |
|---|
| 275 | def write_mislabels(self, final=True): |
|---|
| 276 | if final: |
|---|
| 277 | out_fname = self.mis_fname |
|---|
| 278 | else: |
|---|
| 279 | out_fname = self.premis_fname |
|---|
| 280 | |
|---|
| 281 | with open(out_fname, "w") as fo_all: |
|---|
| 282 | fields = ["SeqID", "MislabeledLevel", "OriginalLabel", "ProposedLabel", "Confidence", "OriginalTaxonomyPath", "ProposedTaxonomyPath", "PerRankConfidence"] |
|---|
| 283 | if self.cfg.ranktest: |
|---|
| 284 | fields += ["HigherRankMisplacedConfidence"] |
|---|
| 285 | header = ";" + "\t".join(fields) + "\n" |
|---|
| 286 | fo_all.write(header) |
|---|
| 287 | if self.cfg.verbose and len(self.mislabels) > 0 and final: |
|---|
| 288 | print "Mislabeled sequences:\n" |
|---|
| 289 | print header |
|---|
| 290 | for mis_rec in self.mislabels: |
|---|
| 291 | output = self.mis_rec_to_string(mis_rec) + "\n" |
|---|
| 292 | fo_all.write(output) |
|---|
| 293 | if self.cfg.verbose and final: |
|---|
| 294 | print(output) |
|---|
| 295 | |
|---|
| 296 | if not final: |
|---|
| 297 | return |
|---|
| 298 | |
|---|
| 299 | if self.cfg.ranktest: |
|---|
| 300 | with open(self.misrank_fname, "w") as fo_all: |
|---|
| 301 | fields = ["RankID", "MislabeledLevel", "OriginalLabel", "ProposedLabel", "Confidence", "OriginalTaxonomyPath", "ProposedTaxonomyPath", "PerRankConfidence"] |
|---|
| 302 | header = ";" + "\t".join(fields) + "\n" |
|---|
| 303 | fo_all.write(header) |
|---|
| 304 | if self.cfg.verbose and len(self.rank_mislabels) > 0: |
|---|
| 305 | print "\nMislabeled higher ranks:\n" |
|---|
| 306 | print header |
|---|
| 307 | for mis_rec in self.rank_mislabels: |
|---|
| 308 | output = self.mis_rec_to_string(mis_rec) + "\n" |
|---|
| 309 | fo_all.write(output) |
|---|
| 310 | if self.cfg.verbose: |
|---|
| 311 | print(output) |
|---|
| 312 | |
|---|
| 313 | self.write_stats() |
|---|
| 314 | |
|---|
| 315 | def run_leave_subtree_out_test(self): |
|---|
| 316 | job_name = self.cfg.subst_name("l1out_rank_%NAME%") |
|---|
| 317 | # if self.jplace_fname: |
|---|
| 318 | # jp = EpaJsonParser(self.jplace_fname) |
|---|
| 319 | # else: |
|---|
| 320 | |
|---|
| 321 | #create file with subtrees |
|---|
| 322 | rank_tips = {} |
|---|
| 323 | rank_parent = {} |
|---|
| 324 | for node in self.tax_tree.traverse("postorder"): |
|---|
| 325 | if node.is_leaf() or node.is_root(): |
|---|
| 326 | continue |
|---|
| 327 | tax_path = node.name |
|---|
| 328 | ranks = Taxonomy.split_rank_uid(tax_path) |
|---|
| 329 | rank_lvl = Taxonomy.lowest_assigned_rank_level(ranks) |
|---|
| 330 | if rank_lvl < 2: |
|---|
| 331 | continue |
|---|
| 332 | |
|---|
| 333 | parent_ranks = Taxonomy.split_rank_uid(node.up.name) |
|---|
| 334 | parent_lvl = Taxonomy.lowest_assigned_rank_level(parent_ranks) |
|---|
| 335 | if parent_lvl < 1: |
|---|
| 336 | continue |
|---|
| 337 | |
|---|
| 338 | rank_seqs = node.get_leaf_names() |
|---|
| 339 | rank_size = len(rank_seqs) |
|---|
| 340 | if rank_size < 2 or rank_size > self.reftree_size-4: |
|---|
| 341 | continue |
|---|
| 342 | |
|---|
| 343 | # print rank_lvl, "\t", tax_path, "\t", rank_seqs, "\n" |
|---|
| 344 | rank_tips[tax_path] = node.get_leaf_names() |
|---|
| 345 | rank_parent[tax_path] = parent_ranks |
|---|
| 346 | |
|---|
| 347 | subtree_list = rank_tips.items() |
|---|
| 348 | |
|---|
| 349 | if len(subtree_list) == 0: |
|---|
| 350 | return 0 |
|---|
| 351 | |
|---|
| 352 | subtree_list_file = self.cfg.tmp_fname("treelist_%NAME%.txt") |
|---|
| 353 | with open(subtree_list_file, "w") as fout: |
|---|
| 354 | for rank_name, tips in subtree_list: |
|---|
| 355 | fout.write("%s\n" % " ".join(tips)) |
|---|
| 356 | |
|---|
| 357 | jp_list = self.raxml.run_epa(job_name, self.refalign_fname, self.reftree_fname, self.optmod_fname, |
|---|
| 358 | mode="l1o_subtree", subtree_fname=subtree_list_file) |
|---|
| 359 | |
|---|
| 360 | subtree_count = 0 |
|---|
| 361 | for jp in jp_list: |
|---|
| 362 | placements = jp.get_placement() |
|---|
| 363 | for place in placements: |
|---|
| 364 | ranks, lws = self.classify_seq(place) |
|---|
| 365 | tax_path = subtree_list[subtree_count][0] |
|---|
| 366 | orig_ranks = Taxonomy.split_rank_uid(tax_path) |
|---|
| 367 | rank_level = Taxonomy.lowest_assigned_rank_level(orig_ranks) |
|---|
| 368 | rank_prefix = self.guess_rank_level_name(orig_ranks, rank_level)[0] |
|---|
| 369 | rank_name = orig_ranks[rank_level] |
|---|
| 370 | if not rank_name.startswith(rank_prefix): |
|---|
| 371 | rank_name = rank_prefix + rank_name |
|---|
| 372 | parent_ranks = rank_parent[tax_path] |
|---|
| 373 | # print orig_ranks, "\n", parent_ranks, "\n", ranks, "\n" |
|---|
| 374 | mis_rec = self.check_rank_tax_labels(rank_name, parent_ranks, ranks, lws) |
|---|
| 375 | if mis_rec: |
|---|
| 376 | self.misrank_conf_map[tax_path] = mis_rec['conf'] |
|---|
| 377 | subtree_count += 1 |
|---|
| 378 | |
|---|
| 379 | return subtree_count |
|---|
| 380 | |
|---|
| 381 | def run_leave_seq_out_test(self): |
|---|
| 382 | job_name = self.cfg.subst_name("l1out_seq_%NAME%") |
|---|
| 383 | placements = [] |
|---|
| 384 | if self.cfg.jplace_fname: |
|---|
| 385 | if os.path.isdir(self.cfg.jplace_fname): |
|---|
| 386 | jplace_fmask = os.path.join(self.cfg.jplace_fname, '*.jplace') |
|---|
| 387 | else: |
|---|
| 388 | jplace_fmask = self.cfg.jplace_fname |
|---|
| 389 | |
|---|
| 390 | jplace_fname_list = glob.glob(jplace_fmask) |
|---|
| 391 | for jplace_fname in jplace_fname_list: |
|---|
| 392 | jp = EpaJsonParser(jplace_fname) |
|---|
| 393 | placements += jp.get_placement() |
|---|
| 394 | |
|---|
| 395 | config.log.debug("Loaded %d placements from %s\n", len(placements), jplace_fmask) |
|---|
| 396 | else: |
|---|
| 397 | jp = self.raxml.run_epa(job_name, self.refalign_fname, self.reftree_fname, self.optmod_fname, mode="l1o_seq") |
|---|
| 398 | placements = jp.get_placement() |
|---|
| 399 | if self.cfg.output_interim_files: |
|---|
| 400 | out_jplace_fname = self.cfg.out_fname("%NAME%.l1out_seq.jplace") |
|---|
| 401 | self.raxml.copy_epa_jplace(job_name, out_jplace_fname, move=True, mode="l1o_seq") |
|---|
| 402 | |
|---|
| 403 | seq_count = 0 |
|---|
| 404 | l1out_ass = {} |
|---|
| 405 | for place in placements: |
|---|
| 406 | seq_name = place["n"][0] |
|---|
| 407 | |
|---|
| 408 | # get original taxonomic label |
|---|
| 409 | # orig_ranks = self.get_orig_ranks(seq_name) |
|---|
| 410 | orig_ranks = self.taxtree_helper.get_seq_ranks_from_tree(seq_name) |
|---|
| 411 | |
|---|
| 412 | # get EPA tax label |
|---|
| 413 | ranks, lws = self.classify_seq(place) |
|---|
| 414 | l1out_ass[seq_name] = (ranks, lws) |
|---|
| 415 | |
|---|
| 416 | # check if they match |
|---|
| 417 | mis_rec = self.check_seq_tax_labels(seq_name, orig_ranks, ranks, lws) |
|---|
| 418 | # cross-check with higher rank mislabels |
|---|
| 419 | if self.cfg.ranktest and mis_rec: |
|---|
| 420 | rank_conf = 0 |
|---|
| 421 | for lvl in range(2,len(orig_ranks)): |
|---|
| 422 | tax_path = Taxonomy.get_rank_uid(orig_ranks, lvl) |
|---|
| 423 | if tax_path in self.misrank_conf_map: |
|---|
| 424 | rank_conf = max(rank_conf, self.misrank_conf_map[tax_path]) |
|---|
| 425 | mis_rec['rank_conf'] = rank_conf |
|---|
| 426 | seq_count += 1 |
|---|
| 427 | |
|---|
| 428 | self.write_assignments(l1out_ass, final=False) |
|---|
| 429 | |
|---|
| 430 | return seq_count |
|---|
| 431 | |
|---|
| 432 | def run_final_epa_test(self): |
|---|
| 433 | self.reftree_outgroup = self.refjson.get_outgroup() |
|---|
| 434 | |
|---|
| 435 | tmp_reftree = self.reftree.copy(method="newick") |
|---|
| 436 | name2refnode = {} |
|---|
| 437 | for leaf in tmp_reftree.iter_leaves(): |
|---|
| 438 | name2refnode[leaf.name] = leaf |
|---|
| 439 | |
|---|
| 440 | tmp_taxtree = self.tax_tree.copy(method="newick") |
|---|
| 441 | name2taxnode = {} |
|---|
| 442 | for leaf in tmp_taxtree.iter_leaves(): |
|---|
| 443 | name2taxnode[leaf.name] = leaf |
|---|
| 444 | |
|---|
| 445 | for mis_rec in self.mislabels: |
|---|
| 446 | rname = mis_rec['name'] |
|---|
| 447 | # rname = EpacConfig.REF_SEQ_PREFIX + name |
|---|
| 448 | |
|---|
| 449 | if rname in name2refnode: |
|---|
| 450 | name2refnode[rname].delete() |
|---|
| 451 | else: |
|---|
| 452 | print "Node not found in the reference tree: %s" % rname |
|---|
| 453 | |
|---|
| 454 | if rname in name2taxnode: |
|---|
| 455 | name2taxnode[rname].delete() |
|---|
| 456 | else: |
|---|
| 457 | print "Node not found in the taxonomic tree: %s" % rname |
|---|
| 458 | |
|---|
| 459 | # remove unifurcation at the root |
|---|
| 460 | if len(tmp_reftree.children) == 1: |
|---|
| 461 | tmp_reftree = tmp_reftree.children[0] |
|---|
| 462 | |
|---|
| 463 | self.mislabels = [] |
|---|
| 464 | |
|---|
| 465 | th = TaxTreeHelper(self.cfg, self.origin_taxonomy) |
|---|
| 466 | th.set_mf_rooted_tree(tmp_taxtree) |
|---|
| 467 | |
|---|
| 468 | epa_result = self.run_epa_once(tmp_reftree) |
|---|
| 469 | |
|---|
| 470 | reftree_epalbl_str = epa_result.get_std_newick_tree() |
|---|
| 471 | placements = epa_result.get_placement() |
|---|
| 472 | |
|---|
| 473 | # update branchid-taxonomy mapping to account for possible changes in branch numbering |
|---|
| 474 | reftree_tax = Tree(reftree_epalbl_str) |
|---|
| 475 | th.set_bf_unrooted_tree(reftree_tax) |
|---|
| 476 | bid_tax_map = th.get_bid_taxonomy_map() |
|---|
| 477 | |
|---|
| 478 | self.write_bid_tax_map(bid_tax_map, final=True) |
|---|
| 479 | |
|---|
| 480 | cl = TaxClassifyHelper(self.cfg, bid_tax_map, self.rate, self.node_height) |
|---|
| 481 | |
|---|
| 482 | # newtax_fname = self.cfg.subst_name("newtax_%NAME%.tre") |
|---|
| 483 | # th.get_tax_tree().write(outfile=newtax_fname, format=3) |
|---|
| 484 | |
|---|
| 485 | final_ass = {} |
|---|
| 486 | for place in placements: |
|---|
| 487 | seq_name = place["n"][0] |
|---|
| 488 | |
|---|
| 489 | # get original taxonomic label |
|---|
| 490 | orig_ranks = self.taxtree_helper.get_seq_ranks_from_tree(seq_name) |
|---|
| 491 | |
|---|
| 492 | # EXPERIMENTAL FEATURE - disabled for now! |
|---|
| 493 | # It could happen that certain ranks were present in the "original" reference tree, but |
|---|
| 494 | # are completely missing in the pruned tree (e.g., all seqs of a species were considered "suspicious" |
|---|
| 495 | # after the leave-one-out test and thus pruned) |
|---|
| 496 | # In this case, EPA has no chance to infer full original taxonomic annotation (=species) since the corresponding clade |
|---|
| 497 | # is now missing. To account for this fact, we amend the original taxonomic annotation and set ranks missing from |
|---|
| 498 | # pruned tree to "Undefined". |
|---|
| 499 | # orig_ranks = th.strip_missing_ranks(orig_ranks) |
|---|
| 500 | # print orig_ranks |
|---|
| 501 | |
|---|
| 502 | # get EPA tax label |
|---|
| 503 | ranks, lws = cl.classify_seq(place["p"]) |
|---|
| 504 | final_ass[seq_name] = (ranks, lws) |
|---|
| 505 | |
|---|
| 506 | #print seq_name, ": ", orig_ranks, "--->", ranks |
|---|
| 507 | |
|---|
| 508 | # check if they match |
|---|
| 509 | mis_rec = self.check_seq_tax_labels(seq_name, orig_ranks, ranks, lws) |
|---|
| 510 | |
|---|
| 511 | self.write_assignments(final_ass, final=True) |
|---|
| 512 | |
|---|
| 513 | def run_epa_once(self, reftree): |
|---|
| 514 | reftree_fname = self.cfg.tmp_fname("final_ref_%NAME%.tre") |
|---|
| 515 | job_name = self.cfg.subst_name("final_epa_%NAME%") |
|---|
| 516 | |
|---|
| 517 | reftree.write(outfile=reftree_fname) |
|---|
| 518 | |
|---|
| 519 | # IMPORTANT: don't load the model, since it's invalid for the pruned true !!! |
|---|
| 520 | optmod_fname="" |
|---|
| 521 | epa_result = self.raxml.run_epa(job_name, self.refalign_fname, reftree_fname, optmod_fname) |
|---|
| 522 | |
|---|
| 523 | if self.cfg.output_interim_files: |
|---|
| 524 | out_jplace_fname = self.cfg.out_fname("%NAME%.final_epa.jplace") |
|---|
| 525 | self.raxml.copy_epa_jplace(job_name, out_jplace_fname, move=True) |
|---|
| 526 | |
|---|
| 527 | return epa_result |
|---|
| 528 | |
|---|
| 529 | def run_test(self): |
|---|
| 530 | self.raxml = RaxmlWrapper(self.cfg) |
|---|
| 531 | |
|---|
| 532 | # config.log.info("Number of sequences in the reference: %d\n", self.reftree_size) |
|---|
| 533 | |
|---|
| 534 | self.refjson.get_raxml_readable_tree(self.reftree_fname) |
|---|
| 535 | self.refalign_fname = self.refjson.get_alignment(self.tmp_refaln) |
|---|
| 536 | self.refjson.get_binary_model(self.optmod_fname) |
|---|
| 537 | |
|---|
| 538 | if self.cfg.ranktest: |
|---|
| 539 | config.log.info("Running the leave-one-rank-out test...\n") |
|---|
| 540 | subtree_count = self.run_leave_subtree_out_test() |
|---|
| 541 | |
|---|
| 542 | config.log.info("Running the leave-one-sequence-out test...\n") |
|---|
| 543 | self.run_leave_seq_out_test() |
|---|
| 544 | |
|---|
| 545 | if len(self.mislabels) > 0: |
|---|
| 546 | config.log.info("Leave-one-out test identified %d suspicious sequences; running final EPA test to check them...\n", len(self.mislabels)) |
|---|
| 547 | if self.cfg.debug: |
|---|
| 548 | self.write_mislabels(final=False) |
|---|
| 549 | self.run_final_epa_test() |
|---|
| 550 | |
|---|
| 551 | self.filter_mislabels() |
|---|
| 552 | self.sort_mislabels() |
|---|
| 553 | self.write_mislabels() |
|---|
| 554 | config.log.info("\nTotal mislabels: %d / %.2f %%", len(self.mislabels), (float(len(self.mislabels)) / self.reftree_size * 100)) |
|---|
| 555 | |
|---|
| 556 | def parse_args(): |
|---|
| 557 | parser = ArgumentParser(usage="%(prog)s -s ALIGNMENT -t TAXONOMY -x {BAC,BOT,ZOO,VIR} [options]", |
|---|
| 558 | description=EpacConfig.SATIVA_INFO % "SATIVA", |
|---|
| 559 | epilog="Example: sativa.py -s example/test.phy -t example/test.tax -x BAC", |
|---|
| 560 | formatter_class=RawDescriptionHelpFormatter) |
|---|
| 561 | parser.add_argument("-s", dest="align_fname", |
|---|
| 562 | help="""Reference alignment file (PHYLIP or FASTA). Sequences must be aligned, |
|---|
| 563 | their IDs must correspond to those in taxonomy file.""") |
|---|
| 564 | parser.add_argument("-t", dest="taxonomy_fname", |
|---|
| 565 | help="""Reference taxonomy file.""") |
|---|
| 566 | parser.add_argument("-x", dest="taxcode_name", choices=["bac", "bot", "zoo", "vir"], type = str.lower, |
|---|
| 567 | help="""Taxonomic code: BAC(teriological), BOT(anical), ZOO(logical), VIR(ological)""") |
|---|
| 568 | parser.add_argument("-n", dest="output_name", default=None, |
|---|
| 569 | help="""Job name, will be used as a prefix for output file names (default: taxonomy file name without extension)""") |
|---|
| 570 | parser.add_argument("-o", dest="output_dir", default=".", |
|---|
| 571 | help="""Output directory (default: current).""") |
|---|
| 572 | parser.add_argument("-T", dest="num_threads", type=int, default=multiprocessing.cpu_count(), |
|---|
| 573 | help="""Specify the number of CPUs (default: %d)""" % multiprocessing.cpu_count()) |
|---|
| 574 | parser.add_argument("-N", dest="rep_num", type=int, default=1, |
|---|
| 575 | help="""Number of RAxML tree searches (with distinct random seeds) to resolve multifurcation. Default: 1""") |
|---|
| 576 | parser.add_argument("-v", dest="verbose", action="store_true", |
|---|
| 577 | help="""Print additional info messages to the console.""") |
|---|
| 578 | parser.add_argument("-R", dest="restart", action="store_true", |
|---|
| 579 | help="""Resume execution after a premature termination (e.g., due to expired job time limit). |
|---|
| 580 | Run name of the previous (terminated) job must be specified via -n option.""") |
|---|
| 581 | parser.add_argument("-c", dest="config_fname", default=None, |
|---|
| 582 | help="Config file name.") |
|---|
| 583 | parser.add_argument("-r", dest="ref_fname", |
|---|
| 584 | help="""Specify the reference alignment and taxonomy in refjson format.""") |
|---|
| 585 | parser.add_argument("-j", dest="jplace_fname", default=None, |
|---|
| 586 | help="""Do not call RAxML EPA, use existing .jplace file as input instead. |
|---|
| 587 | This could be also a directory with *.jplace files.""") |
|---|
| 588 | parser.add_argument("-p", dest="rand_seed", type=int, default=12345, |
|---|
| 589 | help="""Random seed to be used with RAxML. Default: 12345""") |
|---|
| 590 | parser.add_argument("-C", dest="conf_cutoff", type=float, default=0., |
|---|
| 591 | help="""Confidence cut-off between 0 and 1. Default: 0\n""") |
|---|
| 592 | parser.add_argument("-P", dest="brlen_pv", type=float, default=0., |
|---|
| 593 | help="""P-value for branch length Erlang test. Default: 0=off\n""") |
|---|
| 594 | parser.add_argument("-l", dest="min_lhw", type=float, default=0., |
|---|
| 595 | help="""A value between 0 and 1, the minimal sum of likelihood weight of |
|---|
| 596 | an assignment to a specific rank. This value represents a confidence |
|---|
| 597 | measure of the assignment, assignments below this value will be discarded. |
|---|
| 598 | Default: 0 to output all possbile assignments.""") |
|---|
| 599 | parser.add_argument("-m", dest="mfresolv_method", choices=["thorough", "fast", "ultrafast"], |
|---|
| 600 | default="thorough", help="""Method of multifurcation resolution: |
|---|
| 601 | thorough use stardard constrainted RAxML tree search (default) |
|---|
| 602 | fast use RF distance as search convergence criterion (RAxML -D option) |
|---|
| 603 | ultrafast optimize model+branch lengths only (RAxML -f e option)""") |
|---|
| 604 | parser.add_argument("-debug", dest="debug", action="store_true", |
|---|
| 605 | help="""Debug mode, intermediate files will not be cleaned up.""") |
|---|
| 606 | parser.add_argument("-ranktest", dest="ranktest", action="store_true", |
|---|
| 607 | help="""Test for misplaced higher ranks.""") |
|---|
| 608 | parser.add_argument("-tmpdir", dest="temp_dir", default=None, |
|---|
| 609 | help="""Directory for temporary files.""") |
|---|
| 610 | |
|---|
| 611 | args = parser.parse_args() |
|---|
| 612 | if len(sys.argv) == 1: |
|---|
| 613 | parser.print_help() |
|---|
| 614 | sys.exit() |
|---|
| 615 | check_args(args, parser) |
|---|
| 616 | return args |
|---|
| 617 | |
|---|
| 618 | |
|---|
| 619 | def check_args(args, parser): |
|---|
| 620 | if args.ref_fname: |
|---|
| 621 | if args.align_fname: |
|---|
| 622 | print("WARNING: -r and -s options are mutually exclusive! Your alignment file will be ignored.\n") |
|---|
| 623 | if args.taxonomy_fname: |
|---|
| 624 | print("WARNING: -r and -t options are mutually exclusive! Your taxonomy file will be ignored.\n") |
|---|
| 625 | if args.taxcode_name: |
|---|
| 626 | print("WARNING: -r and -x options are mutually exclusive! The taxonomic code from reference file will be used.\n") |
|---|
| 627 | elif not args.align_fname or not args.taxonomy_fname or not args.taxcode_name: |
|---|
| 628 | print("ERROR: either reference in JSON format or taxonomy, alignment and taxonomic code name must be provided:\n") |
|---|
| 629 | parser.print_help() |
|---|
| 630 | sys.exit() |
|---|
| 631 | |
|---|
| 632 | if not os.path.exists(args.output_dir): |
|---|
| 633 | print("Output directory does not exists: %s" % args.output_dir) |
|---|
| 634 | sys.exit() |
|---|
| 635 | |
|---|
| 636 | #check if taxonomy file exists |
|---|
| 637 | if args.taxonomy_fname and not os.path.isfile(args.taxonomy_fname): |
|---|
| 638 | print "ERROR: Taxonomy file not found: %s" % args.taxonomy_fname |
|---|
| 639 | sys.exit() |
|---|
| 640 | |
|---|
| 641 | #check if alignment file exists |
|---|
| 642 | if args.align_fname and not os.path.isfile(args.align_fname): |
|---|
| 643 | print "ERROR: Alignment file not found: %s" % args.align_fname |
|---|
| 644 | sys.exit() |
|---|
| 645 | |
|---|
| 646 | if args.ref_fname and not os.path.isfile(args.ref_fname): |
|---|
| 647 | print("Input reference json file does not exists: %s" % args.ref_fname) |
|---|
| 648 | sys.exit() |
|---|
| 649 | |
|---|
| 650 | if args.jplace_fname and not (os.path.isfile(args.jplace_fname) or os.path.isdir(args.jplace_fname)): |
|---|
| 651 | print("EPA placement file does not exists: %s" % args.jplace_fname) |
|---|
| 652 | sys.exit() |
|---|
| 653 | |
|---|
| 654 | if args.min_lhw < 0 or args.min_lhw > 1.0: |
|---|
| 655 | args.min_lhw = 0.0 |
|---|
| 656 | |
|---|
| 657 | if args.conf_cutoff < 0 or args.conf_cutoff > 1.0: |
|---|
| 658 | args.min_lhw = 0.0 |
|---|
| 659 | |
|---|
| 660 | sativa_home = os.path.dirname(os.path.abspath(__file__)) |
|---|
| 661 | if not args.config_fname: |
|---|
| 662 | args.config_fname = os.path.join(sativa_home, "sativa.cfg") |
|---|
| 663 | if not args.temp_dir: |
|---|
| 664 | args.temp_dir = os.path.join(sativa_home, "tmp") |
|---|
| 665 | if not args.output_name: |
|---|
| 666 | if args.taxonomy_fname: |
|---|
| 667 | base_fname = args.taxonomy_fname |
|---|
| 668 | else: |
|---|
| 669 | base_fname = args.ref_fname |
|---|
| 670 | args.output_name = os.path.splitext(base_fname)[0] |
|---|
| 671 | |
|---|
| 672 | def print_run_info(config): |
|---|
| 673 | print "" |
|---|
| 674 | config.print_version("SATIVA") |
|---|
| 675 | |
|---|
| 676 | call_str = " ".join(sys.argv) |
|---|
| 677 | config.log.info("SATIVA was called as follows:\n\n%s\n" % call_str) |
|---|
| 678 | |
|---|
| 679 | if config.verbose: |
|---|
| 680 | config.log.info("Mislabels search is running with the following parameters:") |
|---|
| 681 | if config.align_fname: |
|---|
| 682 | config.log.info(" Alignment: %s", config.align_fname) |
|---|
| 683 | config.log.info(" Taxonomy: %s", config.taxonomy_fname) |
|---|
| 684 | if config.load_refjson: |
|---|
| 685 | config.log.info(" Reference: %s", config.refjson_fname) |
|---|
| 686 | if config.jplace_fname: |
|---|
| 687 | config.log.info(" EPA jplace file: %s", config.jplace_fname) |
|---|
| 688 | #config.log.info(" Min likelihood weight: %f", args.min_lhw) |
|---|
| 689 | # config.log.info(" Assignment method: %s", args.method) |
|---|
| 690 | config.log.info(" Output directory: %s", os.path.abspath(config.output_dir)) |
|---|
| 691 | config.log.info(" Job name / output files prefix: %s", config.name) |
|---|
| 692 | config.log.info(" Model of rate heterogeneity: %s", config.raxml_model) |
|---|
| 693 | config.log.info(" Confidence cut-off: %f", config.conf_cutoff) |
|---|
| 694 | # config.log.info(" P-value for branch length test: %g", config.brlen_pv) |
|---|
| 695 | config.log.info(" Number of threads: %d", config.num_threads) |
|---|
| 696 | config.log.info("") |
|---|
| 697 | |
|---|
| 698 | if config.debug: |
|---|
| 699 | config.log.debug("Running in DEBUG mode, temp files will be saved to: %s\n", os.path.abspath(config.temp_dir)) |
|---|
| 700 | |
|---|
| 701 | if __name__ == "__main__": |
|---|
| 702 | args = parse_args() |
|---|
| 703 | config = SativaConfig(args) |
|---|
| 704 | |
|---|
| 705 | start_time = time.time() |
|---|
| 706 | trainer_time = 0 |
|---|
| 707 | |
|---|
| 708 | t = LeaveOneTest(config) |
|---|
| 709 | print_run_info(config) |
|---|
| 710 | |
|---|
| 711 | if config.load_refjson: |
|---|
| 712 | t.load_refjson(config.refjson_fname) |
|---|
| 713 | else: |
|---|
| 714 | config.log.info("*** STEP 1: Building the reference tree using provided alignment and taxonomic annotations ***\n") |
|---|
| 715 | tr_start_time = time.time() |
|---|
| 716 | t.run_epa_trainer() |
|---|
| 717 | trainer_time = time.time() - tr_start_time |
|---|
| 718 | t.load_refjson(config.refjson_fname) |
|---|
| 719 | config.log.info("*** STEP 2: Searching for mislabels ***\n") |
|---|
| 720 | |
|---|
| 721 | l1out_start_time = time.time() |
|---|
| 722 | |
|---|
| 723 | t.run_test() |
|---|
| 724 | |
|---|
| 725 | config.clean_tempdir() |
|---|
| 726 | |
|---|
| 727 | l1out_time = time.time() - l1out_start_time |
|---|
| 728 | |
|---|
| 729 | config.log.info("\nResults were saved to: %s", os.path.abspath(t.mis_fname)) |
|---|
| 730 | config.log.info("Execution log was saved to: %s\n", os.path.abspath(config.log_fname)) |
|---|
| 731 | |
|---|
| 732 | elapsed_time = time.time() - start_time |
|---|
| 733 | config.log.info("Analysis completed successfully, elapsed time: %.0f seconds (%.0fs reftree, %.0fs leave-one-out)\n", elapsed_time, trainer_time, l1out_time) |
|---|
