def _tax_diff_table(self, tax1, tax2, output_table):
"""Tabulate incongruency of taxonomy strings at each rank."""
fout = open(output_table, 'w')
fout.write('Lineage\tNo. Extent Taxa')
for rank_label in Taxonomy.rank_labels:
fout.write('\t%s (%%)' % rank_label.title())
fout.write('\n')
taxonomy = Taxonomy()
named_lineages_at_rank = taxonomy.named_lineages_at_rank(tax1)
for rank, taxa in named_lineages_at_rank.iteritems():
rank_label = Taxonomy.rank_labels[rank]
if rank_label == 'species':
continue
extant_taxa_for_rank = taxonomy.extant_taxa_for_rank(rank_label, tax1)
for taxon in taxa:
extent_taxa = extant_taxa_for_rank[taxon]
fout.write('%s\t%d' % (taxon, len(extent_taxa)))
row = defaultdict(list)
for genome_id in extent_taxa:
taxa1 = tax1[genome_id]
taxa2 = tax2[genome_id]
for cur_rank, (taxa1, taxa2) in enumerate(zip(taxa1, taxa2)):
row[cur_rank].append(taxa1 == taxa2)
for cur_rank, matches in row.iteritems():
if cur_rank <= rank:
fout.write('\t-')
else:
perc_match = sum(matches) * 100.0 / len(matches)
fout.write('\t%.1f' % (100.0 - perc_match))
fout.write('\n')
fout.close()
def _tax_diff_table(self, tax1, tax2, output_table):
"""Tabulate incongruency of taxonomy strings at each rank."""
fout = open(output_table, 'w')
fout.write('Lineage\tNo. Extent Taxa')
for rank_label in Taxonomy.rank_labels:
fout.write('\t%s (%%)' % rank_label.title())
fout.write('\n')
taxonomy = Taxonomy()
named_lineages_at_rank = taxonomy.named_lineages_at_rank(tax1)
for rank, taxa in named_lineages_at_rank.items():
rank_label = Taxonomy.rank_labels[rank]
if rank_label == 'species':
continue
extant_taxa_for_rank = taxonomy.extant_taxa_for_rank(rank_label, tax1)
for taxon in taxa:
extent_taxa = extant_taxa_for_rank[taxon]
fout.write('%s\t%d' % (taxon, len(extent_taxa)))
row = defaultdict(list)
for genome_id in extent_taxa:
taxa1 = tax1[genome_id]
taxa2 = tax2[genome_id]
for cur_rank, (taxa1, taxa2) in enumerate(list(zip(taxa1, taxa2))):
row[cur_rank].append(taxa1 == taxa2)
for cur_rank, matches in row.items():
if cur_rank <= rank:
fout.write('\t-')
else:
perc_match = sum(matches) * 100.0 / len(matches)
fout.write('\t%.1f' % (100.0 - perc_match))
fout.write('\n')
fout.close()
def tree_tax_diff(self, tree1_file, tree2_file, output_dir):
"""Tabulate differences between two taxonomies on a tree.
Parameters
----------
tree1_file : str
File with tree in Newick format.
tree2_file : str
File with tree in Newick format.
output_dir : str
Output directory.
"""
tree1 = dendropy.Tree.get_from_path(tree1_file,
schema='newick',
rooting='force-rooted',
preserve_underscores=True)
tree2 = dendropy.Tree.get_from_path(tree2_file,
schema='newick',
rooting='force-rooted',
preserve_underscores=True)
# prune both trees to a set of common taxa
taxa1 = set()
for t in tree1.leaf_node_iter():
taxa1.add(t.taxon.label)
taxa2 = set()
for t in tree2.leaf_node_iter():
taxa2.add(t.taxon.label)
taxa_in_common = taxa1.intersection(taxa2)
self.logger.info('Tree 1 contains %d taxa.' % len(taxa1))
self.logger.info('Tree 2 contains %d taxa.' % len(taxa2))
self.logger.info('Pruning trees to the %d taxa in common.' % len(taxa_in_common))
tree1.retain_taxa_with_labels(taxa_in_common)
tree2.retain_taxa_with_labels(taxa_in_common)
# get named lineages at each taxonomic rank
taxonomy = Taxonomy()
tax1 = taxonomy.read_from_tree(tree1)
tax2 = taxonomy.read_from_tree(tree2)
taxa_at_rank1 = taxonomy.named_lineages_at_rank(tax1)
taxa_at_rank2 = taxonomy.named_lineages_at_rank(tax2)
# identify retained taxonomic names
tax_file_name = os.path.splitext(os.path.basename(tree1_file))[0]
output_file = os.path.join(output_dir, '%s.taxa_diff.tsv' % tax_file_name)
fout = open(output_file, 'w')
fout.write('Rank\tClassification\tTaxonomy 1\tTaxonomy 2\n')
taxon2_accounted_for = defaultdict(set)
for rank, rank_label in enumerate(Taxonomy.rank_labels[0:-1]):
for taxon in taxa_at_rank1[rank]:
# check if taxon has been retained
if taxon in taxa_at_rank2[rank]:
fout.write('%s\t%s\t%s\t%s\n' % (rank_label, 'retained', taxon, taxon))
taxon2_accounted_for[rank].add(taxon)
continue
# check if name was simply corrected by changing suffix
old_taxon = self._change_suffix(taxon, rank, taxa_at_rank2)
if old_taxon:
fout.write('%s\t%s\t%s\t%s\n' % (rank_label, 'corrected', taxon, old_taxon))
taxon2_accounted_for[rank].add(old_taxon)
continue
# check if taxon has been moved up or down in rank
old_taxon, old_rank = self._renamed(taxon, rank, taxa_at_rank2)
if old_taxon:
if rank < old_rank:
fout.write('%s\t%s\t%s\t%s\n' % (rank_label, 'more general', taxon, old_taxon))
elif rank == old_rank:
fout.write('%s\t%s\t%s\t%s\n' % (rank_label, 'corrected', taxon, old_taxon))
else:
fout.write('%s\t%s\t%s\t%s\n' % (rank_label, 'more specific', taxon, old_taxon))
taxon2_accounted_for[old_rank].add(old_taxon)
continue
# otherwise, the taxon appears to be new
fout.write('%s\t%s\t%s\t%s\n' % (rank_label, 'new', taxon, 'NA'))
# report deprecated taxa
for rank, rank_label in enumerate(Taxonomy.rank_labels[0:-1]):
for taxon in taxa_at_rank2[rank]:
if taxon not in taxon2_accounted_for[rank]:
fout.write('%s\t%s\t%s\t%s\n' % (rank_label, 'deprecated', 'NA', taxon))
fout.close()
# tabulate congruence of taxonomy strings
output_table = os.path.join(output_dir, '%s.perc_diff.tsv' % tax_file_name)
self._tax_diff_table(tax1, tax2, output_table)
def tree_tax_diff(self, tree1_file, tree2_file, output_dir):
"""Tabulate differences between two taxonomies on a tree.
Parameters
----------
tree1_file : str
File with tree in Newick format.
tree2_file : str
File with tree in Newick format.
output_dir : str
Output directory.
"""
tree1 = dendropy.Tree.get_from_path(tree1_file,
schema='newick',
rooting='force-rooted',
preserve_underscores=True)
tree2 = dendropy.Tree.get_from_path(tree2_file,
schema='newick',
rooting='force-rooted',
preserve_underscores=True)
# prune both trees to a set of common taxa
taxa1 = set()
for t in tree1.leaf_node_iter():
taxa1.add(t.taxon.label)
taxa2 = set()
for t in tree2.leaf_node_iter():
taxa2.add(t.taxon.label)
taxa_in_common = taxa1.intersection(taxa2)
self.logger.info('Tree 1 contains %d taxa.' % len(taxa1))
self.logger.info('Tree 2 contains %d taxa.' % len(taxa2))
self.logger.info('Pruning trees to the %d taxa in common.' % len(taxa_in_common))
tree1.retain_taxa_with_labels(taxa_in_common)
tree2.retain_taxa_with_labels(taxa_in_common)
# get named lineages at each taxonomic rank
taxonomy = Taxonomy()
tax1 = taxonomy.read_from_tree(tree1)
tax2 = taxonomy.read_from_tree(tree2)
taxa_at_rank1 = taxonomy.named_lineages_at_rank(tax1)
taxa_at_rank2 = taxonomy.named_lineages_at_rank(tax2)
# identify retained taxonomic names
tax_file_name = os.path.splitext(os.path.basename(tree1_file))[0]
output_file = os.path.join(output_dir, '%s.taxa_diff.tsv' % tax_file_name)
fout = open(output_file, 'w')
fout.write('Rank\tClassification\tTaxonomy 1\tTaxonomy 2\n')
taxon2_accounted_for = defaultdict(set)
for rank, rank_label in enumerate(Taxonomy.rank_labels[0:-1]):
for taxon in taxa_at_rank1[rank]:
# check if taxon has been retained
if taxon in taxa_at_rank2[rank]:
fout.write('%s\t%s\t%s\t%s\n' % (rank_label, 'retained', taxon, taxon))
taxon2_accounted_for[rank].add(taxon)
continue
# check if name was simply corrected by changing suffix
old_taxon = self._change_suffix(taxon, rank, taxa_at_rank2)
if old_taxon:
fout.write('%s\t%s\t%s\t%s\n' % (rank_label, 'corrected', taxon, old_taxon))
taxon2_accounted_for[rank].add(old_taxon)
continue
# check if taxon has been moved up or down in rank
old_taxon, old_rank = self._renamed(taxon, rank, taxa_at_rank2)
if old_taxon:
if rank < old_rank:
fout.write('%s\t%s\t%s\t%s\n' % (rank_label, 'more general', taxon, old_taxon))
elif rank == old_rank:
fout.write('%s\t%s\t%s\t%s\n' % (rank_label, 'corrected', taxon, old_taxon))
else:
fout.write('%s\t%s\t%s\t%s\n' % (rank_label, 'more specific', taxon, old_taxon))
taxon2_accounted_for[old_rank].add(old_taxon)
continue
# otherwise, the taxon appears to be new
fout.write('%s\t%s\t%s\t%s\n' % (rank_label, 'new', taxon, 'NA'))
# report deprecated taxa
for rank, rank_label in enumerate(Taxonomy.rank_labels[0:-1]):
for taxon in taxa_at_rank2[rank]:
if taxon not in taxon2_accounted_for[rank]:
fout.write('%s\t%s\t%s\t%s\n' % (rank_label, 'deprecated', 'NA', taxon))
fout.close()
# tabulate congruence of taxonomy strings
output_table = os.path.join(output_dir, '%s.perc_diff.tsv' % tax_file_name)
self._tax_diff_table(tax1, tax2, output_table)
