def cluster_contigs_anvio(self):
# clustering of contigs is done for each configuration file under static/clusterconfigs/merged directory;
# at this point we don't care what those recipes really require because we already merged and generated
# every data file that may be required.
self.run.info_single("Anvi'o hierarchical clustering of contigs...", nl_before=1, nl_after=1, mc="blue")
if not self.skip_hierarchical_clustering:
for config_name in self.clustering_configs:
config_path = self.clustering_configs[config_name]
config = ClusteringConfiguration(config_path, self.output_directory, db_paths=self.database_paths, row_ids_of_interest=self.split_names)
try:
clustering_id, newick = clustering.order_contigs_simple(config, distance=self.distance, linkage=self.linkage, progress=self.progress)
except Exception as e:
self.run.warning('Clustering has failed for "%s": "%s"' % (config_name, e))
self.progress.end()
continue
_, distance, linkage = clustering_id.split(':')
dbops.add_items_order_to_db(anvio_db_path=self.profile_db_path,
order_name=config_name,
order_data=newick,
distance=distance,
linkage=linkage,
make_default=config_name == constants.merged_default,
run=self.run)
def cluster_contigs_anvio(self):
# clustering of contigs is done for each configuration file under static/clusterconfigs/merged directory;
# at this point we don't care what those recipes really require because we already merged and generated
# every data file that may be required.
self.run.info_single("Anvi'o hierarchical clustering of contigs...", nl_before=1, nl_after=1, mc="blue")
if not self.skip_hierarchical_clustering:
for config_name in self.clustering_configs:
config_path = self.clustering_configs[config_name]
config = ClusteringConfiguration(config_path, self.output_directory, db_paths=self.database_paths, row_ids_of_interest=self.split_names)
try:
clustering_id, newick = clustering.order_contigs_simple(config, distance=self.distance, linkage=self.linkage, progress=self.progress)
except Exception as e:
self.run.warning('Clustering has failed for "%s": "%s"' % (config_name, e))
self.progress.end()
continue
_, distance, linkage = clustering_id.split(':')
dbops.add_items_order_to_db(anvio_db_path=self.profile_db_path,
order_name=config_name,
order_data=newick,
distance=distance,
linkage=linkage,
make_default=config_name == constants.merged_default,
run=self.run)
def gen_synteny_based_ordering_of_gene_clusters(self, gene_clusters_dict):
"""Take the dictionary of gene_clusters, and order gene_clusters per genome based on synteny of genes.
This adds more orders to the pangenomic output so the user can enforce ordering of
gene_clusters based on the synteny of genes they contain in a given genome.
The synteny in this context is defined by the gene caller ids. Gene caller ids
follow a numerical order in anvi'o contigs databases for genes that are coming
from the same contig. Of course, the synteny does not mean much for genes that
fragmented into multiple contigs.
"""
# yes. this is meren converting the gene_clusters_dict into a pandas data frame :/ if you are reading
# this line and if you are not evan, don't tell evan about this. everyone else: i don't know
# what you're talking about.
df = pd.DataFrame(
list(chain.from_iterable(list(gene_clusters_dict.values()))))
df = df.sort_values(by=['genome_name', 'gene_caller_id'])
df = df.reset_index(drop=True)
# forced synteny
for genome_name in df.genome_name.unique():
gene_clusters_in_genome = df.loc[(
df.genome_name == genome_name)].gene_cluster_id.unique()
gene_clusters_not_described = df.loc[~df.gene_cluster_id.isin(
gene_clusters_in_genome)].gene_cluster_id.unique()
gene_clusters_order_based_on_genome_synteny = list(
gene_clusters_in_genome) + list(gene_clusters_not_described)
order_name = 'Forced synteny <> %s' % genome_name
dbops.add_items_order_to_db(
self.pan_db_path,
order_name,
','.join(gene_clusters_order_based_on_genome_synteny),
order_data_type_newick=False,
run=self.run)
gene_cluster_gene_cluster_edges = []
# network description of gene_cluster-gene_cluster relationships given the gene synteny.
gene_ordered_list_of_gene_clusters = list(
zip(df.gene_caller_id, df.gene_cluster_id))
for index in range(1, len(gene_ordered_list_of_gene_clusters)):
(GENE_A, gene_cluster_A), (
GENE_B, gene_cluster_B) = gene_ordered_list_of_gene_clusters[
index - 1], gene_ordered_list_of_gene_clusters[index]
if GENE_A == GENE_B - 1:
gene_cluster_gene_cluster_edges.append(
(gene_cluster_A, gene_cluster_B), )
def gen_synteny_based_ordering_of_gene_clusters(self, gene_clusters_dict):
"""Take the dictionary of gene_clusters, and order gene_clusters per genome based on synteny of genes.
This adds more orders to the pangenomic output so the user can enforce ordering of
gene_clusters based on the synteny of genes they contain in a given genome.
The synteny in this context is defined by the gene caller ids. Gene caller ids
follow a numerical order in anvi'o contigs databases for genes that are coming
from the same contig. Of course, the synteny does not mean much for genmes that
fragmented into multiple contigs.
"""
# yes. this is meren converting the gene_clusters_dict into a pandas data frame :/ if you are reading
# this line and if you are not evan, don't tell evan about this. everyone else: i don't know
# what you're talking about.
df = pd.DataFrame(list(chain.from_iterable(list(gene_clusters_dict.values()))))
df = df.sort_values(by=['genome_name', 'gene_caller_id'])
df = df.reset_index(drop=True)
# forced synteny
for genome_name in df.genome_name.unique():
gene_clusters_in_genome = df.loc[(df.genome_name == genome_name)].gene_cluster_id.unique()
gene_clusters_not_described = df.loc[~df.gene_cluster_id.isin(gene_clusters_in_genome)].gene_cluster_id.unique()
gene_clusters_order_based_on_genome_synteny = list(gene_clusters_in_genome) + list(gene_clusters_not_described)
order_name = 'Forced synteny <> %s' % genome_name
dbops.add_items_order_to_db(self.pan_db_path, order_name, ','.join(gene_clusters_order_based_on_genome_synteny), order_data_type_newick=False, run=self.run)
gene_cluster_gene_cluster_edges = []
# network description of gene_cluster-gene_cluster relationshops given the gene synteny.
gene_ordered_list_of_gene_clusters = list(zip(df.gene_caller_id, df.gene_cluster_id))
for index in range(1, len(gene_ordered_list_of_gene_clusters)):
(GENE_A, gene_cluster_A), (GENE_B, gene_cluster_B) = gene_ordered_list_of_gene_clusters[index-1], gene_ordered_list_of_gene_clusters[index]
if GENE_A == GENE_B - 1:
gene_cluster_gene_cluster_edges.append((gene_cluster_A, gene_cluster_B), )
