# Copyright 2019 Damien Goutte-Gattat. All rights reserved. # # This file is part of the Biopython distribution and governed by your # choice of the "Biopython License Agreement" or the "BSD 3-Clause License". # Please see the LICENSE file that should have been included as part of this # package. """Bio.SeqIO support for the "gck" file format. The GCK binary format is generated by the Gene Construction Kit software from Textco BioSoftware, Inc. """ from struct import unpack from Bio import Alphabet from Bio._utils import _read_header from Bio.File import as_handle from Bio.Seq import Seq from Bio.SeqFeature import SeqFeature, FeatureLocation from Bio.SeqRecord import SeqRecord def _read(handle, length): """Read the specified number of bytes from the given handle.""" data = handle.read(length) if len(data) < length: raise ValueError("Cannot read {} bytes from handle".format(length)) return data def _read_packet(handle): """Read a length-prefixed packet. Parts of a GCK file are made of "packets" comprising of 4 bytes giving the packet's size, followed by the packet's data. There is no type tag. The type of a packet, and thus the type of data it contains, is solely indicated by the position of the packet within the GCK file. """ length = _read(handle, 4) length = unpack(">I", length)[0] data = _read(handle, length) return (data, length) def _read_pstring(handle): """Read a Pascal string. A Pascal string is one byte for length followed by the actual string. """ length = _read(handle, 1) length = unpack(">B", length)[0] data = _read(handle, length).decode("ASCII") return data def _read_p4string(handle): """Read a 32-bit Pascal string. Similar to a Pascal string but length is encoded on 4 bytes. """ length = _read(handle, 4) length = unpack(">I", length)[0] data = _read(handle, length).decode("ASCII") return data def GckIterator(handle): """Parse a GCK file and return a SeqRecord object. Note that a GCK file can only contain one sequence, so this iterator will always return a single record. """ with as_handle(handle, "rb") as fp: # Skip file header # GCK files start with a 24-bytes header. Bytes 4 and 8 seem to # always be 12, maybe this could act as a magic cookie. Bytes # 17-20 and 21-24 contain variable values of unknown meaning. # check if file is empty _read_header(handle, 24) # Read the actual sequence data packet, length = _read_packet(handle) # The body of the sequence packet starts with a 32-bit integer # representing the length of the sequence. seq_length = unpack(">I", packet[:4])[0] # This length should not be larger than the length of the # sequence packet. if seq_length > length - 4: raise ValueError("Conflicting sequence length values") sequence = packet[4:].decode("ASCII") record = SeqRecord(Seq(sequence, alphabet=Alphabet.generic_dna)) # Skip unknown packet _read_packet(handle) # Read features packet packet, length = _read_packet(handle) (seq_length, num_features) = unpack(">IH", packet[:6]) # Check that length in the features packet matches the actual # length of the sequence if seq_length != len(record): raise ValueError("Conflicting sequence length values") # Each feature is stored in a 92-bytes structure. if length - 6 != num_features * 92: raise ValueError( "Features packet size inconsistent with number of features" ) for i in range(0, num_features): offset = 6 + i * 92 feature_data = packet[offset : offset + 92] # There's probably more stuff to unpack in that structure, # but those values are the only ones I understand. (start, end, type, strand, has_name, has_comment, version) = unpack( ">II6xH14xB17xII35xB", feature_data ) if strand == 1: # Reverse strand strand = -1 else: # Other possible values are 0 (no strand specified), # 2 (forward strand), and 3 (both strands). All are # treated as a forward strand. strand = 1 location = FeatureLocation(start, end, strand=strand) # It looks like any value > 0 indicates a CDS... if type > 0: type = "CDS" else: type = "misc_feature" # Each feature may have a name and a comment, which are then # stored immediately after the features packet. Names are # stored as Pascal strings (1 length byte followed by the # string itself), comments are stored as "32-bit Pascal strings" # (4 length bytes followed by the string). qualifiers = {} if has_name > 0: name = _read_pstring(handle) qualifiers["label"] = [name] if has_comment > 0: comment = _read_p4string(handle) qualifiers["note"] = [comment] # Each feature may exist in several "versions". We keep only # the most recent version. if version > 0: continue feature = SeqFeature(location, type=type, qualifiers=qualifiers) record.features.append(feature) # Read restriction sites packet # We are not interested in restriction sites, but we must still read # that packet so that we can skip the names and comments for each # site, which are stored after that packet in a similar way as for # the features above. packet, length = _read_packet(handle) (seq_length, num_sites) = unpack(">IH", packet[:6]) # Each site is stored in a 88-bytes structure if length - 6 != num_sites * 88: raise ValueError("Sites packet size inconsistent with number of sites") for i in range(0, num_sites): offset = 6 + i * 88 site_data = packet[offset : offset + 88] (start, end, has_name, has_comment) = unpack(">II24xII48x", site_data) # Skip names and comments if has_name: _read_pstring(handle) if has_comment: _read_p4string(handle) # Skip unknown packet _read_packet(handle) # Next in the file are "version packets". # However they are not properly formatted "packets" as they are not # preceded by an integer giving their size. Instead we have a # short integer indicating how many versions are there, and then # as many 260-bytes block as we have versions. num_versions = _read(handle, 2) num_versions = unpack(">H", num_versions)[0] versions = _read(handle, num_versions * 260) for i in range(0, num_versions): offset = i * 260 version_data = versions[offset : offset + 260] # Each version may have a comment, which is then stored # after all the "version packets". has_comment = unpack(">I", version_data[-4:])[0] if has_comment > 0: _read_p4string(handle) # Skip unknown fixed-size block # Whatever this block contains, it is not preceded by any length # indicator, so I hope its size is indeed constant in all files... _read(handle, 706) # Read the construct's name name = _read_pstring(handle) record.name = record.id = name.split(" ")[0] record.description = name # Circularity byte # There may be other flags in that block, but their meaning # is unknown to me. flags = _read(handle, 17) circularity = unpack(">16xB", flags)[0] if circularity > 0: record.annotations["topology"] = "circular" else: record.annotations["topology"] = "linear" yield record