def getGeneSequencesFromFastas(pangenome, fasta_file):
fastaDict = {}
for line in read_compressed_or_not(fasta_file):
elements = [el.strip() for el in line.split("\t")]
if len(elements) <= 1:
logging.getLogger().error(
"No tabulation separator found in organisms file")
exit(1)
org = pangenome.addOrganism(elements[0])
with read_compressed_or_not(elements[1]) as currFastaFile:
fastaDict[org] = read_fasta(org, currFastaFile)
if not set(pangenome.organisms) <= set(fastaDict.keys()):
missing = len(pangenome.organisms) - len(
set(pangenome.organisms) & set(fastaDict.keys()))
raise Exception(
f"Not all of your pangenome's organisms are present within the provided fasta file. {missing} are missing (out of {len(pangenome.organisms)})."
)
for org in pangenome.organisms:
try:
for contig in org.contigs:
for gene in contig.genes:
gene.add_dna(
get_dna_sequence(fastaDict[org][contig.name], gene))
for rna in contig.RNAs:
rna.add_dna(
get_dna_sequence(fastaDict[org][contig.name], gene))
except KeyError:
msg = f"Fasta file for organism {org.name} did not have the contig {contig.name} that was read from the annotation file. "
msg += f"The provided contigs in the fasta were : { ', '.join([contig for contig in fastaDict[org].keys()])}."
raise KeyError(msg)
pangenome.status["geneSequences"] = "Computed"
def read_org_gff(organism,
gff_file_path,
circular_contigs,
getSeq,
pseudo=False):
(GFF_seqname, _,
GFF_type, GFF_start, GFF_end, _, GFF_strand, _, GFF_attribute) = range(
0, 9) #missing values : source, score, frame. They are unused.
def getGffAttributes(gff_fields):
"""
Parses the gff attribute's line and outputs the attributes in a dict structure.
:param gff_fields: a gff line stored as a list. Each element of the list is a column of the gff.
:type list:
:return: attributes:
:rtype: dict
"""
attributes_field = [
f for f in gff_fields[GFF_attribute].strip().split(';')
if len(f) > 0
]
attributes = {}
for att in attributes_field:
try:
(key, value) = att.strip().split('=')
attributes[key.upper()] = value
except ValueError:
pass #we assume that it is a strange, but useless field for our analysis
return attributes
def getIDAttribute(attributes):
"""
Gets the ID of the element from which the provided attributes were extracted. Raises an error if no ID is found.
:param attribute:
:type dict:
:return: ElementID:
:rtype: string
"""
ElementID = attributes.get("ID")
if not ElementID:
logging.getLogger().error(
"Each CDS type of the gff files must own a unique ID attribute. Not the case for file: "
+ gff_file_path)
exit(1)
return ElementID
hasFasta = False
fastaString = ""
org = Organism(organism)
geneCounter = 0
rnaCounter = 0
with read_compressed_or_not(gff_file_path) as gff_file:
for line in gff_file:
if hasFasta:
fastaString += line
continue
elif line.startswith('##', 0, 2):
if line.startswith('FASTA', 2, 7):
if not getSeq: #if getting the sequences is useless...
break
hasFasta = True
elif line.startswith('sequence-region', 2, 17):
fields = [el.strip() for el in line.split()]
contig = org.getOrAddContig(
fields[1],
True if fields[1] in circular_contigs else False)
continue
elif line.startswith(
'#!', 0, 2
): ## special refseq comment lines for versionning softs, assemblies and annotations.
continue
gff_fields = [el.strip() for el in line.split('\t')]
attributes = getGffAttributes(gff_fields)
pseudogene = False
if gff_fields[GFF_type] == 'region':
if gff_fields[GFF_seqname] in circular_contigs:
contig.is_circular = True
elif gff_fields[GFF_type] == 'CDS' or "RNA" in gff_fields[GFF_type]:
geneID = attributes.get(
"PROTEIN_ID"
) #if there is a 'PROTEIN_ID' attribute, it's where the ncbi stores the actual gene ids, so we use that.
if geneID is None: #if its not found, we get the one under the 'ID' field which must exist (otherwise not a gff3 compliant file)
geneID = getIDAttribute(attributes)
try:
name = attributes.pop('NAME')
except KeyError:
try:
name = attributes.pop('GENE')
except KeyError:
name = ""
if "pseudo" in attributes or "pseudogene" in attributes:
pseudogene = True
try:
product = attributes.pop('PRODUCT')
except KeyError:
product = ""
try:
genetic_code = attributes.pop("TRANSL_TABLE")
except KeyError:
genetic_code = "11"
if contig.name != gff_fields[GFF_seqname]:
contig = org.getOrAddContig(
gff_fields[GFF_seqname]) #get the current contig
if gff_fields[GFF_type] == "CDS" and (not pseudogene or
(pseudogene and pseudo)):
gene = Gene(org.name + "_CDS_" + str(geneCounter).zfill(4))
#here contig is filled in order, so position is the number of genes already stored in the contig.
gene.fill_annotations(start=int(gff_fields[GFF_start]),
stop=int(gff_fields[GFF_end]),
strand=gff_fields[GFF_strand],
geneType=gff_fields[GFF_type],
position=len(contig.genes),
name=name,
product=product,
genetic_code=genetic_code,
local_identifier=geneID)
gene.fill_parents(org, contig)
contig.addGene(gene)
geneCounter += 1
elif "RNA" in gff_fields[GFF_type]:
rna = RNA(org.name + "_CDS_" + str(rnaCounter).zfill(4))
rna.fill_annotations(start=int(gff_fields[GFF_start]),
stop=int(gff_fields[GFF_end]),
strand=gff_fields[GFF_strand],
geneType=gff_fields[GFF_type],
name=name,
product=product,
local_identifier=geneID)
rna.fill_parents(org, contig)
contig.addRNA(rna)
rnaCounter += 1
### GET THE FASTA SEQUENCES OF THE GENES
if hasFasta and fastaString != "":
contigSequences = read_fasta(org, fastaString.split('\n'))
for contig in org.contigs:
for gene in contig.genes:
gene.add_dna(
get_dna_sequence(contigSequences[contig.name], gene))
for rna in contig.RNAs:
rna.add_dna(get_dna_sequence(contigSequences[contig.name],
rna))
return org, hasFasta