def check_snps():
""" Check that accessions are sorted """
last_ref = None
for snp in utility.parse_file(args['in']):
last_ref = snp['site_id'].rsplit('|')[0]
for snp in utility.parse_file(args['in']):
ref_id = snp['site_id'].rsplit('|')[0]
if ref_id < last_snp['ref_id']:
sys.exit("Accessions not sorted")
else:
last_ref = ref_id
def check_snps():
""" Check that accessions are sorted """
last_ref = None
for snp in utility.parse_file(args['in']):
last_ref = snp['site_id'].rsplit('|')[0]
for snp in utility.parse_file(args['in']):
ref_id = snp['site_id'].rsplit('|')[0]
if ref_id < last_snp['ref_id']:
sys.exit("Accessions not sorted")
else:
last_ref = ref_id
def open_snp_info(indir):
""" return generator for snps info file """
inpath = '%s/snps_info.txt' % indir
if not os.path.isfile(inpath):
return None
else:
return utility.parse_file(inpath)
def open_infiles(species_id, samples):
""" Open SNP files for species across samples """
infiles = []
for sample in samples:
inpath = '%s/snps/output/%s.snps.gz' % (sample.dir, species_id)
infiles.append(utility.parse_file(inpath))
return infiles
def genes_summary(args):
""" Get summary of mapping statistics """
# store stats
stats = {}
inpath = '%s/%s' % (args['outdir'], 'genes/temp/pangenome.map')
for species_id in set(utility.read_ref_to_cluster(inpath).values()):
pangenome_size, covered_genes, total_coverage, marker_coverage = [0,0,0,0]
for r in utility.parse_file('/'.join([args['outdir'], 'genes/output/%s.genes.gz' % species_id])):
pangenome_size += 1
coverage = float(r['coverage'])
normcov = float(r['copy_number'])
if coverage > 0:
covered_genes += 1
total_coverage += coverage
if normcov > 0:
marker_coverage = coverage/normcov
stats[species_id] = {'pangenome_size':pangenome_size,
'covered_genes':covered_genes,
'fraction_covered':covered_genes/float(pangenome_size),
'mean_coverage':total_coverage/covered_genes if covered_genes > 0 else 0.0,
'marker_coverage':marker_coverage}
# write stats
fields = ['pangenome_size', 'covered_genes', 'fraction_covered', 'mean_coverage', 'marker_coverage']
outfile = open('/'.join([args['outdir'], 'genes/summary.txt']), 'w')
outfile.write('\t'.join(['species_id'] + fields)+'\n')
for species_id in stats:
record = [species_id] + [str(stats[species_id][field]) for field in fields]
outfile.write('\t'.join(record)+'\n')
def build_gene_matrices(species_id, samples, args):
""" Compute gene copy numbers for samples """
gene_to_family = read_gene_map(species_id, args)
count_genes = len(gene_to_family.keys())
count_genomes = len(
set(['.'.join(x.split('.')[0:2]) for x in gene_to_family]))
count_families = len(set(gene_to_family.values()))
print(" %s genes from %s genomes" % (count_genes, count_genomes))
print(" clustered into %s families at %s percent id" %
(count_families, args['cluster_pid']))
for sample in samples:
sample.genes = {}
for type in ['presabs', 'copynum', 'depth']:
sample.genes[type] = defaultdict(float)
inpath = '%s/genes/output/%s.genes.gz' % (sample.dir, species_id)
for r in utility.parse_file(inpath):
if 'ref_id' in r:
r['gene_id'] = r['ref_id'] # fix old fields if present
if 'normalized_coverage' in r:
r['copy_number'] = r['normalized_coverage']
if 'raw_coverage' in r: r['coverage'] = r['raw_coverage']
gene_id = gene_to_family[r['gene_id']]
sample.genes['copynum'][gene_id] += float(r['copy_number'])
sample.genes['depth'][gene_id] += float(r['coverage'])
for sample in samples:
for gene_id, copynum in sample.genes['copynum'].items():
if copynum >= args['min_copy']:
sample.genes['presabs'][gene_id] = 1
else:
sample.genes['presabs'][gene_id] = 0
def genes_summary(args):
""" Get summary of mapping statistics """
# store stats
stats = {}
inpath = '%s/%s' % (args['outdir'], 'genes/temp/pangenome.map')
for species_id in set(utility.read_ref_to_cluster(inpath).values()):
pangenome_size, covered_genes, total_coverage, marker_coverage = [0,0,0,0]
for r in utility.parse_file('/'.join([args['outdir'], 'genes/output/%s.genes.gz' % species_id])):
pangenome_size += 1
coverage = float(r['coverage'])
normcov = float(r['copy_number'])
if coverage > 0:
covered_genes += 1
total_coverage += coverage
if normcov > 0:
marker_coverage = coverage/normcov
stats[species_id] = {'pangenome_size':pangenome_size,
'covered_genes':covered_genes,
'fraction_covered':covered_genes/float(pangenome_size),
'mean_coverage':total_coverage/covered_genes if covered_genes > 0 else 0.0,
'marker_coverage':marker_coverage}
# write stats
fields = ['pangenome_size', 'covered_genes', 'fraction_covered', 'mean_coverage', 'marker_coverage']
outfile = open('/'.join([args['outdir'], 'genes/summary.txt']), 'w')
outfile.write('\t'.join(['species_id'] + fields)+'\n')
for species_id in stats:
record = [species_id] + [str(stats[species_id][field]) for field in fields]
outfile.write('\t'.join(record)+'\n')
def snps_summary(args):
""" Get summary of mapping statistics """
# store stats
stats = {}
inpath = os.path.join(args['outdir'], 'snps/temp/genomes.map')
ref_to_species = utility.read_ref_to_cluster(inpath)
for species_id in set(ref_to_species.values()):
genome_length, covered_bases, total_depth, identity, maf = [0,0,0,0,0]
for r in utility.parse_file('/'.join([args['outdir'], 'snps/output/%s.snps.gz' % species_id])):
genome_length += 1
depth = int(r['depth'])
if depth > 0:
covered_bases += 1
total_depth += depth
if r['ref_allele'] == r['cons_allele']:
identity += 1
ref_freq = float(r['ref_freq'])
maf += ref_freq if ref_freq <= 0.5 else 1 - ref_freq
stats[species_id] = {'genome_length':genome_length,
'covered_bases': covered_bases,
'fraction_covered':covered_bases/float(genome_length),
'mean_coverage':total_depth/float(covered_bases) if covered_bases > 0 else 0
}
# write stats
fields = ['genome_length', 'covered_bases', 'fraction_covered', 'mean_coverage']
outfile = open('/'.join([args['outdir'], 'snps/summary.txt']), 'w')
outfile.write('\t'.join(['species_id'] + fields)+'\n')
for species_id in stats:
record = [species_id] + [str(stats[species_id][field]) for field in fields]
outfile.write('\t'.join(record)+'\n')
def open_infiles(species_id, samples): """ Open SNP files for species across samples """ infiles = [] for sample in samples: inpath = '%s/snps/output/%s.snps.gz' % (sample.dir, species_id) infiles.append(utility.parse_file(inpath)) return infiles
def read_annotations(args):
annotations = {}
inpath = '%s/species_info.txt' % args['db']
if not os.path.isfile(inpath): sys.exit("File not found: %s" % inpath)
for rec in utility.parse_file(inpath):
annotations[rec['species_id']] = rec['species_name']
return annotations
def read_annotations(args):
annotations = {}
inpath = '%s/species_info.txt' % args['db']
if not os.path.isfile(inpath): sys.exit("File not found: %s" % inpath)
for rec in utility.parse_file(inpath):
annotations[rec['species_id']] = rec['species_name']
return annotations
def add_summary_stats(args, samples): """ add summary stats """ for rec in utility.parse_file(args['summary']): if 'run_accession' in rec: rec['sample_id'] = rec['run_accession'] if rec['sample_id'] in samples: samples[rec['sample_id']].mean_depth = float(rec['average_depth']) samples[rec['sample_id']].fraction_covered = float(rec['fraction_covered']) samples['between'].mean_depth = np.mean([_.mean_depth for _ in samples.values() if _.id != 'between'])
def read_cluster_map(sp, db, pid):
sp.map = {}
for ext in ['', '.gz']:
path = '/'.join([db, 'pan_genomes', sp.id, 'gene_info.txt%s' % ext])
if os.path.isfile(path):
sp.gene_info = path
for r in utility.parse_file(sp.gene_info):
sp.map[r['centroid_99']] = r['centroid_%s' % pid]
def read_stats(inpath, type):
stats = {}
for rec in utility.parse_file(inpath):
if 'cluster_id' in rec: rec['species_id'] = rec['cluster_id']
if 'phyeco_coverage' in rec: rec['marker_coverage'] = rec['phyeco_coverage']
if 'average_depth' in rec: rec['mean_coverage'] = rec['average_depth']
if type=='genes':
rec['fraction_covered'] = float(rec['covered_genes'])/float(rec['pangenome_size'])
stats[rec['species_id']] = rec
return stats
def read_marker_info(args):
""" Read info for marker genes from phyeco.fa """
info = {}
for seq in Bio.SeqIO.parse('%s/marker_genes/phyeco.fa' % args['db'],
'fasta'):
info[seq.id] = None
for r in utility.parse_file('%s/marker_genes/phyeco.map' % args['db']):
if r['gene_id'] in info:
info[r['gene_id']] = r
return info
def read_stats(inpath, type):
stats = {}
for rec in utility.parse_file(inpath):
if 'cluster_id' in rec: rec['species_id'] = rec['cluster_id']
if 'phyeco_coverage' in rec: rec['marker_coverage'] = rec['phyeco_coverage']
if 'average_depth' in rec: rec['mean_coverage'] = rec['average_depth']
if type=='genes':
rec['fraction_covered'] = float(rec['covered_genes'])/float(rec['pangenome_size'])
stats[rec['species_id']] = rec
return stats
def fetch_samples(args):
""" initialize samples from indir and set pass_qc flag """
samples = []
count_hq = 0
stop = False
filters = {'min_depth':args['sample_depth'], 'min_cov':args['fract_cov']}
for info in utility.parse_file('%s/snps_summary.txt' % args['indir']):
samples.append(Sample(info, filters, stop))
if samples[-1].pass_qc: count_hq += 1
if count_hq >= args['max_samples']: stop = True
return samples
def read_genes(db, species_id, contigs):
""" Read in gene coordinates from features file """
genes_path = '%s/rep_genomes/%s/genome.features.gz' % (db, species_id)
genes = []
for gene in utility.parse_file(genes_path):
if gene['gene_type'] == 'RNA':
continue
else:
gene['start'] = int(gene['start'])
gene['end'] = int(gene['end'])
gene['seq'] = get_gene_seq(gene, contigs[gene['scaffold_id']])
genes.append(gene)
return genes
def read_abundance(inpath):
""" Parse species abundance file """
if not os.path.isfile(inpath):
sys.exit("\nCould not locate species profile: %s\nTry rerunning with run_midas.py species" % inpath)
abun = {}
for rec in utility.parse_file(inpath):
# format record
if 'cluster_id' in rec: rec['species_id'] = rec['cluster_id']
if 'count_reads' in rec: rec['count_reads'] = int(rec['count_reads'])
if 'coverage' in rec: rec['coverage'] = float(rec['coverage'])
if 'relative_abundance' in rec: rec['relative_abundance'] = float(rec['relative_abundance'])
abun[rec['species_id']] = rec
return abun
def call_markers(species, samples, args):
""" determine if marker present in each sample """
# open marker list
markers = utility.parse_file(species.paths['markers'])
marker = fetch_marker(markers) # dictionary for 1st marker allele
if marker is None:
sys.exit("\nError: no marker alleles found in file: %s\n" %
species.paths['markers'])
# init markers per sample
for sample in samples.values():
sample.markers = set([])
# loop over sites
sites = parse.fetch_sites(species, samples)
for index, site in enumerate(sites):
# record progress
if not index % 100000: print("%s sites processed" % index)
# stop early
if index >= args['max_sites']: break
# skip sites not in marker list
if (site.ref_id != marker['ref_id']
or site.ref_pos < marker['ref_pos']):
continue
# determine if marker present in each sample
for sample in site.samples.values():
# skip samples without marker
if sample.depth == 0:
continue
elif marker['allele'] == site.ref_allele:
sample.marker_freq = sample.ref_freq
elif marker['allele'] == sample.alt_allele:
sample.marker_freq = 1 - sample.ref_freq
else:
continue
# record marker allele for sample_id
sample.marker_count = round(sample.marker_freq * sample.depth)
if (sample.marker_freq >= args['min_freq']
and sample.marker_count >= args['min_reads']):
sample.markers.add(site.id)
# fetch next marker allele
marker = fetch_marker(markers)
if marker is None: break
def read_genes(db, species_id, contigs):
""" Read in gene coordinates from features file """
genes_path = '%s/genome_clusters/%s/genome.features.gz' % (db, species_id)
genes = []
for gene in utility.parse_file(genes_path):
if gene['gene_type'] == 'rna':
continue
else:
gene['accession'] = 'accn|%s' % gene['accession']
gene['start'] = int(gene['start'])
gene['end'] = int(gene['end'])
gene['seq'] = get_gene_seq(gene, contigs[gene['accession']])
genes.append(gene)
return genes
def read_genes(db, species_id, contigs): """ Read in gene coordinates from features file """ genes_path = '%s/genome_clusters/%s/genome.features.gz' % (db, species_id) genes = [] for gene in utility.parse_file(genes_path): if gene['gene_type'] == 'rna': continue else: gene['accession'] = 'accn|%s' % gene['accession'] gene['start'] = int(gene['start']) gene['end'] = int(gene['end']) gene['seq'] = get_gene_seq(gene, contigs[gene['accession']]) genes.append(gene) return genes
def select_species(args, type='genes'):
""" Select all species with a minimum number of high-coverage samples"""
# read species annotations
species_info = {}
inpath = os.path.join(args['db'], 'species_info.txt')
for rec in utility.parse_file(inpath):
species_info[rec['species_id']] = rec
# fetch all species with at least 1 sample
species = {}
for sample in load_samples(args):
if not sample.paths[type]:
sys.stderr.write("Warning: no %s output for sample: %s\n" %
(type, sample.dir))
continue
for id, info in read_stats(sample.paths[type], type).items():
if (args['species_id']
and id not in args['species_id'].split(',')):
continue # skip unspecified species
elif (args['max_samples'] and id in species
and len(species[id].samples) >= args['max_samples']):
continue # skip species with too many samples
elif float(info['mean_coverage']) < args['sample_depth']:
continue # skip low-coverage sample
elif type == 'snps' and float(
info['fraction_covered']) < args['fract_cov']:
continue # skip low-coverage sample
if id not in species:
species[id] = Species(id,
species_info) # initialize new species
species[id].samples.append(sample) # append sample
# dict to list
species = list(species.values())
# remove species with an insufficient number of samples
species = [
sp for sp in species if len(sp.samples) >= int(args['min_samples'])
]
# sort by number of samples
species = sort_species(species)
# select a subset of species to analyze
if args['max_species'] is not None and len(species) > args['max_species']:
species = species[0:args['max_species']]
# create species directories
for sp in species:
outdir = os.path.join(args['outdir'], sp.id)
if not os.path.isdir(outdir): os.mkdir(outdir)
print(" found %s species with sufficient high-coverage samples\n" %
len(species))
return species
def parse_sites(indir, samples):
""" yield genomic sites from input files """
index = 0
files = {} # open input files
for ext in ['alt_allele', 'depth', 'ref_freq']:
files[ext] = parse_tsv('%s/snps_%s.txt' % (indir, ext))
info = utility.parse_file('%s/snps_info.txt' % indir)
while True: # yield GenomicSite
site = GenomicSite(files, samples, info)
if not site.id:
break
else:
index += 1
yield site
for file in files.values(): # close input files
file.close()
def read_abundance(inpath):
""" Parse species abundance file """
if not os.path.isfile(inpath):
sys.exit(
"\nCould not locate species profile: %s\nTry rerunning with run_midas.py species"
% inpath)
abun = {}
for rec in utility.parse_file(inpath):
# format record
if 'species_id' in rec: rec['species_id'] = rec['species_id']
if 'count_reads' in rec: rec['count_reads'] = int(rec['count_reads'])
if 'coverage' in rec: rec['coverage'] = float(rec['coverage'])
if 'relative_abundance' in rec:
rec['relative_abundance'] = float(rec['relative_abundance'])
abun[rec['species_id']] = rec
return abun
def parse_sites(indir, site_depth=0, info_path=None, max_sites=None):
""" Parse reference frequency matrix and snp info file """
index = 0
sample_ids = list_sample_ids('%s/snps_ref_freq.txt' % indir)
freq = open('%s/snps_ref_freq.txt' % indir); next(freq)
depth = open('%s/snps_depth.txt' % indir); next(depth)
alleles = open('%s/snps_alt_allele.txt' % indir); next(alleles)
info = utility.parse_file(info_path) if info_path else None
while True:
site = GenomicSite(sample_ids, freq, depth, alleles, site_depth, info)
if not site.id:
break
elif max_sites and index >= max_sites:
break
else:
index += 1
yield site
freq.close(); depth.close()
def select_species(args, type='genes'):
""" Select all species with a minimum number of high-coverage samples"""
# read species annotations
species_info = {}
inpath = os.path.join(args['db'], 'species_info.txt')
for rec in utility.parse_file(inpath):
species_info[rec['species_id']] = rec
# fetch all species with at least 1 sample
species = {}
for sample in load_samples(args):
if not sample.paths[type]:
sys.stderr.write("Warning: no genes output for sample: %s\n" % sample.dir)
continue
for id, info in read_stats(sample.paths[type], type).items():
if (args['species_id']
and id not in args['species_id'].split(',')):
continue # skip unspecified species
elif (args['max_samples']
and id in species
and len(species[id].samples) >= args['max_samples']):
continue # skip species with too many samples
elif float(info['mean_coverage']) < args['sample_depth']:
continue # skip low-coverage sample
elif type=='snps' and float(info['fraction_covered']) < args['fract_cov']:
continue # skip low-coverage sample
if id not in species:
species[id] = Species(id, species_info) # initialize new species
species[id].samples.append(sample) # append sample
# dict to list
species = species.values()
# remove species with an insufficient number of samples
species = [sp for sp in species if len(sp.samples) >= args['min_samples']]
# sort by number of samples
species = sort_species(species)
# select a subset of species to analyze
if args['max_species'] is not None and len(species) > args['max_species']:
species = species[0:args['max_species']]
# create species directories
for sp in species:
outdir = os.path.join(args['outdir'], sp.id)
if not os.path.isdir(outdir): os.mkdir(outdir)
print " found %s species with sufficient high-coverage samples\n" % len(species)
return species
def snps_summary(args):
""" Get summary of mapping statistics """
# store stats
stats = {}
inpath = os.path.join(args['outdir'], 'snps/temp/genomes.map')
ref_to_species = utility.read_ref_to_cluster(inpath)
for species_id in set(ref_to_species.values()):
genome_length, covered_bases, total_depth, identity, maf = [
0, 0, 0, 0, 0
]
for r in utility.parse_file('/'.join(
[args['outdir'],
'snps/output/%s.snps.gz' % species_id])):
genome_length += 1
depth = int(r['depth'])
if depth > 0:
covered_bases += 1
total_depth += depth
if r['ref_allele'] == r['cons_allele']:
identity += 1
ref_freq = float(r['ref_freq'])
maf += ref_freq if ref_freq <= 0.5 else 1 - ref_freq
stats[species_id] = {
'genome_length':
genome_length,
'covered_bases':
covered_bases,
'fraction_covered':
covered_bases / float(genome_length),
'mean_coverage':
total_depth / float(covered_bases) if covered_bases > 0 else 0
}
# write stats
fields = [
'genome_length', 'covered_bases', 'fraction_covered', 'mean_coverage'
]
outfile = open('/'.join([args['outdir'], 'snps/summary.txt']), 'w')
outfile.write('\t'.join(['species_id'] + fields) + '\n')
for species_id in stats:
record = [species_id
] + [str(stats[species_id][field]) for field in fields]
outfile.write('\t'.join(record) + '\n')
def fetch_samples(args):
""" List and select samples from input """
from midas.analyze.diversity import Sample
samples = []
for info in parse_file('%s/snps_summary.txt' % args['indir']):
# init sample
sample = Sample(info)
sample.pass_qc = True
# keep/exclude sample
if sample.filter(args['sample_depth'], args['fract_cov']):
sample.pass_qc = False
if args['keep_samples'] and sample.id not in args['keep_samples']:
sample.pass_qc = False
if args['exclude_samples'] and sample.id in args['exclude_samples']:
sample.pass_qc = False
if sum([1 for s in samples if s.pass_qc]) >= args['max_samples']:
sample.pass_qc = False
# store sample
samples.append(sample)
# select random sample
if args['rand_samples']: resample_samples(samples)
return samples
def build_gene_matrices(sp, min_copy):
""" Compute gene copy numbers for samples """
for sample in sp.samples:
sample.genes = {}
for field, dtype in [('presabs', float), ('copynum', float),
('depth', float), ('reads', int)]:
sample.genes[field] = defaultdict(dtype)
inpath = '%s/genes/output/%s.genes.gz' % (sample.dir, sp.id)
for r in utility.parse_file(inpath):
if 'ref_id' in r:
r['gene_id'] = r['ref_id'] # fix old fields if present
if 'normalized_coverage' in r:
r['copy_number'] = r['normalized_coverage']
if 'raw_coverage' in r: r['coverage'] = r['raw_coverage']
gene_id = sp.map[r['gene_id']]
sample.genes['copynum'][gene_id] += float(r['copy_number'])
sample.genes['depth'][gene_id] += float(r['coverage'])
sample.genes['reads'][gene_id] += int(
r['count_reads']) if 'count_reads' in r else 0
for sample in sp.samples:
for gene_id, copynum in sample.genes['copynum'].items():
if copynum >= min_copy: sample.genes['presabs'][gene_id] = 1
else: sample.genes['presabs'][gene_id] = 0
def build_gene_matrices(species_id, samples, args):
""" Compute gene copy numbers for samples """
gene_to_family = read_gene_map(species_id, args)
count_genes = len(gene_to_family.keys())
count_genomes = len(set(['.'.join(x.split('.')[0:2]) for x in gene_to_family]))
count_families = len(set(gene_to_family.values()))
print(" %s genes from %s genomes" % (count_genes, count_genomes))
print(" clustered into %s families at %s percent id" % (count_families, args['cluster_pid']))
for sample in samples:
sample.genes = {}
for type in ['presabs', 'copynum', 'depth']:
sample.genes[type] = defaultdict(float)
inpath = '%s/genes/output/%s.genes.gz' % (sample.dir, species_id)
for r in utility.parse_file(inpath):
if 'normalized_coverage' in r: r['copy_number'] = r['normalized_coverage']
if 'raw_coverage' in r: r['coverage'] = r['raw_coverage']
gene_id = gene_to_family[r['gene_id']]
sample.genes['copynum'][gene_id] += float(r['copy_number'])
sample.genes['depth'][gene_id] += float(r['coverage'])
for sample in samples:
for gene_id, copynum in sample.genes['copynum'].items():
if copynum >= args['min_copy']: sample.genes['presabs'][gene_id] = 1
else: sample.genes['presabs'][gene_id] = 0
def snps_summary(args, species):
""" Get summary of mapping statistics """
# store stats
stats = {}
for species_id in species:
genome_length, covered_bases, total_depth, maf = [0, 0, 0, 0]
for r in utility.parse_file('/'.join(
[args['outdir'],
'snps/output/%s.snps.gz' % species_id])):
genome_length += 1
depth = int(r['depth'])
if depth > 0:
covered_bases += 1
total_depth += depth
ref_freq = float(r['ref_freq'])
maf += ref_freq if ref_freq <= 0.5 else 1 - ref_freq
fraction_covered = covered_bases / float(genome_length)
mean_coverage = total_depth / float(
covered_bases) if covered_bases > 0 else 0
stats[species_id] = {
'genome_length': genome_length,
'covered_bases': covered_bases,
'fraction_covered': fraction_covered,
'mean_coverage': mean_coverage
}
# write stats
fields = [
'genome_length', 'covered_bases', 'fraction_covered', 'mean_coverage'
]
outfile = open('/'.join([args['outdir'], 'snps/summary.txt']), 'w')
outfile.write('\t'.join(['species_id'] + fields) + '\n')
for species_id in stats:
record = [species_id
] + [str(stats[species_id][field]) for field in fields]
outfile.write('\t'.join(record) + '\n')
outfile.close()
def list_genes(args):
""" List the set of genes from snps_info.txt """
genes = set([])
for r in parse_file('%s/snps_info.txt' % args['indir']):
if r['gene_id'] != '': genes.add(r['gene_id'])
return genes
def read_annotations(args):
info = {}
inpath = '%s/species_info.txt' % args['db']
for r in utility.parse_file(inpath):
info[r['species_id']] = r
return info
