Documentation

GenbankWriter
in package
implements GenbankWriterInterface

Covers the subset of the GenBank flat-file format GenbankPlasmidMapper reads back : LOCUS, DEFINITION, ACCESSION, a FEATURES table (a synthetic "source 1..length" line plus one entry per PlasmidFeature) and an ORIGIN sequence block. It does not attempt REFERENCE, COMMENT, VERSION, a real division code or submission date - Plasmid carries none of those, and inventing plausible- looking values for them would misrepresent the record rather than describe it. The LOCUS line follows NCBI's fixed columns (see writeLocusLine()), except for a locus name longer than 16 characters, written in full rather than truncated.

A REVERSE-strand feature is written as "complement(start..end)" with start <= end, the standard GenBank convention. A feature crossing the origin (start > end) is written as INSDC does on a circular molecule, "join(start..length,1..end)", wrapped in complement() on the REVERSE strand - the form GenbankPlasmidMapper reads back as the same origin-crossing feature.

DEFINITION and qualifier lines wrap at 79 characters, as NCBI writes them. The source feature carries /organism and /mol_type when the plasmid's metadata gives them (GenbankPlasmidMapper keeps those of the record it read).

A feature with no /gene, /label or /product metadata is written with its name as /label.

INSDC has no notation for an unstranded feature : a location without complement() lies on the direct strand. A Strand::NONE feature is therefore written as a plain range followed by /biophp_strand="none" (UNSTRANDED_QUALIFIER), a qualifier of this library's own, which GenbankPlasmidMapper reads back as Strand::NONE. Any other reader ignores it and, as INSDC says, reads the feature on the direct strand.

A CDS phase is written as /codon_start = phase + 1, a bare number as INSDC specifies. A PROMOTER or TERMINATOR with no GenBank key of its own is written as "regulatory" with the matching /regulatory_class, the "promoter" and "terminator" keys being deprecated since 15-DEC-2014. A feature's GenBank key and its /gene, /label, /product, /note qualifiers are read back from getMetadata()/getNote() exactly as GenbankPlasmidMapper wrote them when importing, so a read-then-write round trip of a feature that originated from GenBank is lossless. A feature with no such metadata (hand-built, imported from GFF3/BED...) falls back to a FeatureType-derived key ; FeatureType's richer vocabulary (INSERT, MARKER, REPORTER, TAG, RESTRICTION_SITE...) has no GenBank feature-key equivalent for most of its members, so those honestly fall back to "misc_feature" rather than a plausible-looking but wrong guess. Class GenbankWriter

Tags
author

Amélie DUVERNET aka Amelaye amelieonline@gmail.com

Table of Contents

Interfaces

GenbankWriterInterface
Interface GenbankWriterInterface - serializes a Plasmid into GenBank flat-file text. The counterpart of GenbankPlasmidMapper, not of ParseGenbankManager directly : it writes a Plasmid aggregate, not the raw Feature entities the legacy parser produces.

Constants

UNSTRANDED_QUALIFIER  : mixed = "biophp_strand"
The qualifier marking a feature on no particular strand, which INSDC cannot write
BASES_PER_GROUP  : mixed = 10
BASES_PER_LINE  : mixed = 60
GENBANK_KEY_BY_FEATURE_TYPE  : array<string, string> = [\Amelaye\BioPHP\Domain\Cloning\ValueObject\Fea...
MAX_LINE_LENGTH  : mixed = 79
REGULATORY_CLASS_BY_FEATURE_TYPE  : array<string, string> = [\Amelaye\BioPHP\Domain\Cloning\ValueObject\Fea...
The /regulatory_class a "regulatory" feature built from these types is written with, INSDC having deprecated the "promoter" and "terminator" keys on 15-DEC-2014.

Methods

write()  : string
locationSignature()  : string
wrap()  : string
Writes a text over as many lines of at most 79 characters as it needs, the first opening on $sPrefix, the next ones indented as deep. Lines break at blanks ; a word longer than a whole line is cut.
writeFeature()  : string
writeLocusLine()  : string
NCBI's fixed LOCUS columns (1-based) : name 13-28, length right-justified 30-40, "bp" 42-43, molecule type 48-53, topology 56-63. A name longer than the 16 columns it is given is written in full, the remaining fields separated by spaces, as NCBI itself does for long locus names ; ParseGenbankManager::parseLocus() reads both back, word by word.
writeOriginBlock()  : string
writeQualifier()  : string

Constants

UNSTRANDED_QUALIFIER

The qualifier marking a feature on no particular strand, which INSDC cannot write

public mixed UNSTRANDED_QUALIFIER = "biophp_strand"

GENBANK_KEY_BY_FEATURE_TYPE

private array<string, string> GENBANK_KEY_BY_FEATURE_TYPE = [\Amelaye\BioPHP\Domain\Cloning\ValueObject\FeatureType::CDS => "CDS", \Amelaye\BioPHP\Domain\Cloning\ValueObject\FeatureType::PROMOTER => "regulatory", \Amelaye\BioPHP\Domain\Cloning\ValueObject\FeatureType::TERMINATOR => "regulatory", \Amelaye\BioPHP\Domain\Cloning\ValueObject\FeatureType::ORIGIN_OF_REPLICATION => "rep_origin"]

REGULATORY_CLASS_BY_FEATURE_TYPE

The /regulatory_class a "regulatory" feature built from these types is written with, INSDC having deprecated the "promoter" and "terminator" keys on 15-DEC-2014.

private array<string, string> REGULATORY_CLASS_BY_FEATURE_TYPE = [\Amelaye\BioPHP\Domain\Cloning\ValueObject\FeatureType::PROMOTER => "promoter", \Amelaye\BioPHP\Domain\Cloning\ValueObject\FeatureType::TERMINATOR => "terminator"]

Methods

locationSignature()

private locationSignature(PlasmidFeature $oFeature) : string
Parameters
$oFeature : PlasmidFeature
Return values
string —

What a feature shares with another one that the reader cannot tell it from

wrap()

Writes a text over as many lines of at most 79 characters as it needs, the first opening on $sPrefix, the next ones indented as deep. Lines break at blanks ; a word longer than a whole line is cut.

private wrap(string $sPrefix, string $sText) : string
Parameters
$sPrefix : string

The label or indentation opening the first line

$sText : string
Return values
string

writeFeature()

private writeFeature(PlasmidFeature $oFeature, int $iLength[, bool $bSharesItsLocation = false ]) : string
Parameters
$oFeature : PlasmidFeature
$iLength : int

The plasmid length, where an origin-crossing feature wraps

$bSharesItsLocation : bool = false

Another feature of the same key lies next to it, at the same place

Return values
string

writeLocusLine()

NCBI's fixed LOCUS columns (1-based) : name 13-28, length right-justified 30-40, "bp" 42-43, molecule type 48-53, topology 56-63. A name longer than the 16 columns it is given is written in full, the remaining fields separated by spaces, as NCBI itself does for long locus names ; ParseGenbankManager::parseLocus() reads both back, word by word.

private writeLocusLine(string $sName, int $iLength) : string
Parameters
$sName : string
$iLength : int
Return values
string

writeOriginBlock()

private writeOriginBlock(string $sSequence) : string
Parameters
$sSequence : string
Return values
string

writeQualifier()

private writeQualifier(string $sQualifier, string $sValue) : string
Parameters
$sQualifier : string
$sValue : string
Return values
string
On this page

Search results