def __init__(self, dbPath, ccdsPath=None, debug=False):
super(Importer, self).__init__()
self.dbPath = dbPath
self.reference(ccdsPath)
# This is where the elements are stored before begin persisted
self.genes = {}
self.txs = {}
self.exons = {}
# Set up the Element Adapter
self.adapter = ElementAdapter(dbPath, debug=debug)
class Importer(object):
"""
Setup class that takes a CCDS database file and pushes it through to form the
basis of a SQLite database. Should take only a few minutes to complete.
:param str dbPath: Full path to the new database
:param str ccdsPath: Path to location of the CCDS file (Default: ``None``,
optional)
:param bool debug: Debug info is printed to the console (Def. `False`,
optional)
"""
def __init__(self, dbPath, ccdsPath=None, debug=False):
super(Importer, self).__init__()
self.dbPath = dbPath
self.reference(ccdsPath)
# This is where the elements are stored before begin persisted
self.genes = {}
self.txs = {}
self.exons = {}
# Set up the Element Adapter
self.adapter = ElementAdapter(dbPath, debug=debug)
def reference(self, ccdsPath):
"""
Public: Sets the path to the CCDS reference file
:param str ccdsPath: Path to the CCDS file
:returns: ``self`` for chainability
"""
self.ccdsPath = ccdsPath
return self
def populate(self):
"""
Public: Where it all happens. Populates the new database with data from the
CCDS file.
I think this is "ACID compliant". It only sets up things if everything
works. I other words. If you have a working database after running this
method you can be sure all elements were successfully added to the
database.
"""
# Setup the new database
self.adapter.setup()
self.reader = csv.reader(open(self.ccdsPath, "r"), delimiter="\t")
for row in self.reader:
# Skip some rows that don't contain useful information
commentLine = row[0].startswith("#")
unofficial = (row[5] != "Public")
if commentLine or unofficial:
continue
(chrom, hgnc, tx_id, strand, tx_start,
tx_end, exonCoords) = self._extractLineData(row)
# Collect all transcript related information
txKwargs = {
"tx_id": tx_id,
"chrom": chrom,
"start": tx_start,
"end": tx_end,
"strand": strand,
"gene_id": hgnc
}
# Check if the gene has already been created
gene = self.genes.get(hgnc, None)
if gene is None:
# Collect all gene related information
gsKwargs = {
"hgnc": hgnc,
"chrom": chrom,
"start": tx_start,
"end": tx_end,
"strand": strand
}
gene = self.genes[hgnc] = self.adapter.create("gene", **gsKwargs)
else:
# Make sure we assign the correct gene coordinates (most extreme)
if gene.start > tx_start:
gene.start = tx_start
if gene.end < tx_end:
gene.end = tx_end
tx = self.txs.get(tx_id, None)
if tx is None:
# Create a new transcript object
tx = self.txs[tx_id] = self.adapter.create("transcript", **txKwargs)
# Add the parent gene to the transcript
tx.gene = gene
# Create new exons
for coord in exonCoords:
ex_id = "{chrom}-{start}-{end}".format(chrom=chrom, start=coord[0],
end=coord[1])
if ex_id in self.exons:
exon = self.exons[ex_id]
else:
# Collect all exon information
exKwargs = {
"ex_id": ex_id,
"chrom": chrom,
"start": coord[0],
"end": coord[1],
"strand": strand
}
exon = self.exons[ex_id] = self.adapter.create("exon", **exKwargs)
# Add the gene and transcript parents to the exon
if not gene in exon.genes:
exon.genes.extend([gene])
if not tx in exon.transcripts:
exon.transcripts.extend([tx])
# Add elements to session
self.adapter.add(self.exons.values())
self.adapter.add(self.genes.values())
self.adapter.add(self.txs.values())
self.adapter.commit()
def _extractLineData(self, row):
"""
Private: Extracts the useful information from one CCDS file row.
:param list row: A list of strings from split row
:returns: A bunch of strings and ints (see order below)
"""
# Chrom, HGNC, Transcript ID, strand, Transcript start, end, exon coords
return (row[0], row[2], row[4], row[6], int(row[7])-1, int(row[8])-1,
self._generateExonCoordinates(row[9]))
def _generateExonCoordinates(self, rowData):
"""
Private: Takes the formatted string of exon coordinates from the CCDS row
and turns it into a more managable list of lists with int start, end
coordinates for each exon.
:param str rowData: A csv string of (start,end) pairs
:returns: A list of lists with the start, end pairs (int)
"""
# Remove the "[]"
csvExons = rowData[1:-1].replace(" ", "")
# 1. Split first into exons coordinates
# 2. Split into start, end and parse int
exons = [[int(pos) for pos in item.split("-")]
for item in csvExons.split(",")]
# 3. Correct coords to 0,0-based Pythonic standard
for exon in exons:
exon[0] -= 1
exon[1] -= 1
return exons
