Get misc header info from PDB header. More...

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "pdb.h"
#include "seq.h"
#include "macros.h"
#include "fsscanf.h"
#include "general.h"
#include "hash.h"

Macros
#define	MAXWORD 8

#define	MAXBUFF 160

#define	ALLOCSIZE 80

#define	CLEAR_BIOMOL(b)

Functions
BOOL	blGetHeaderWholePDB (WHOLEPDB wpdb, char header, int maxheader, char date, int maxdate, char pdbcode, int maxcode)

char *	blGetTitleWholePDB (WHOLEPDB *wpdb)

BOOL	blGetCompoundWholePDBChain (WHOLEPDB wpdb, char chain, COMPND *compnd)

int	blFindMolID (WHOLEPDB wpdb, char chain)

BOOL	blGetSpeciesWholePDBChain (WHOLEPDB wpdb, char chain, PDBSOURCE *source)

BOOL	blGetCompoundWholePDBMolID (WHOLEPDB wpdb, int molid, COMPND compnd)

BOOL	blGetSpeciesWholePDBMolID (WHOLEPDB wpdb, int molid, PDBSOURCE source)

char *	blGetSeqresAsStringWholePDB (WHOLEPDB wpdb, char chains, MODRES modres, BOOL doNucleic)

MODRES *	blGetModresWholePDB (WHOLEPDB *wpdb)

void	blFindOriginalResType (char modAA, char stdAA, MODRES *modres)

BIOMOLECULE *	blGetBiomoleculeWholePDB (WHOLEPDB *wpdb)

void	blFreeBiomolecule (BIOMOLECULE *biomolecule)

HASHTABLE *	blGetSeqresByChainWholePDB (WHOLEPDB wpdb, MODRES modres, BOOL doNucleic)

Detailed Description

Get misc header info from PDB header.

Version: V1.8

Date: 03.10.16

Copyright: (c) Dr. Andrew C.R. Martin / UCL, 2015-2016

Author: Dr. Andrew C. R. Martin

: Institute of Structural & Molecular Biology, University College London, Gower Street, London. WC1E 6BT.

: andre.nosp@m.w@bi.nosp@m.oinf..nosp@m.org..nosp@m.uk andre.nosp@m.w.ma.nosp@m.rtin@.nosp@m.ucl..nosp@m.ac.uk

This code is NOT IN THE PUBLIC DOMAIN, but it may be copied according to the conditions laid out in the accompanying file COPYING.DOC.

The code may be modified as required, but any modifications must be documented so that the person responsible can be identified.

The code may not be sold commercially or included as part of a commercial product except as described in the file COPYING.DOC.

Description:

Usage:

See documentation for details

Revision History:

V1.0 26.03.15 Original
V1.1 13.05.15 blGetTitleWholePDB() returns NULL if no title line in header. Fix for blGetSpeciesWholePDBChain() Added blGetCompoundWholePDBMolID() and blGetSpeciesWholePDBMolID() By: CTP
V1.2 04.06.15 Fixed bug in dealing with compounds where the referenced chains span more than one line
V1.3 09.06.15 Merged changes from CTP and ACRM. Updated blGetTitleWholePDB() By: CTP
V1.4 11.06.15 Added blGetSeqresAsStringWholePDB(), blGetModresWholePDB() and blFindOriginalResType() By: ACRM
V1.5 26.06.15 Added blGetBiomoleculeWholePDB() and blFreeBiomolecule()
V1.6 25.11.15 Added blGetSeqresByChainWholePDB()
V1.7 02.12.15 Sequences were not being terminated properly in blGetSeqresByChainWholePDB()
V1.8 03.10.16 Added <stdlib.h>

Definition in file PDBHeaderInfo.c.

Macro Definition Documentation

#define ALLOCSIZE 80

Definition at line 137 of file PDBHeaderInfo.c.

#define CLEAR_BIOMOL ( b )

Value:

do{                                          \
   (b)->details         = NULL;                 \
   (b)->numBiomolecules = 0;                    \
   (b)->biomolNumber    = 0;                    \
   (b)->authorUnit[0]   = '\0';                 \
   (b)->softwareUnit[0] = '\0';                 \
   (b)->chains          = NULL;                 \
   (b)->biomt           = NULL;                 \
   }  while(0)

Definition at line 139 of file PDBHeaderInfo.c.

#define MAXBUFF 160

Definition at line 136 of file PDBHeaderInfo.c.

#define MAXWORD 8

Definition at line 135 of file PDBHeaderInfo.c.

Function Documentation

int blFindMolID	(	WHOLEPDB *	wpdb,
		char *	chain
	)

Parameters

[in]	*wpdb	WHOLEPDB structure
[in]	*chain	Chain label

Returns: MOL_ID or 0 if chain not found

Finds the MOL_ID for a specified chain

28.04.15 Original By: ACRM
04.06.15 Modified to use the ExtractField() routine instead of duplicating code here. Fixes a bug in dealing with compounds where the referenced chains span more than one line

Definition at line 481 of file PDBHeaderInfo.c.

void blFindOriginalResType	(	char *	modAA,
		char *	stdAA,
		MODRES *	modres
	)

Parameters

[in]	*modAA	Non-standard (modified) amino acid name
[out]	*stdAA	Standard amino acid from which it was derived
[in]	*modres	MODRES linked list

Uses the MODRES information to identify the original (standard) amino acid from which a modified amino acid was derived

07.03.07 Original By: ACRM
11.06.15 Moved to bioplib, renamed routine and parameters

Definition at line 940 of file PDBHeaderInfo.c.

void blFreeBiomolecule ( BIOMOLECULE * biomolecule )

Definition at line 1252 of file PDBHeaderInfo.c.

BIOMOLECULE* blGetBiomoleculeWholePDB ( WHOLEPDB * wpdb )

Parameters

[in] *wpdb Pointer to WHOLEPDB structure

Returns: Pointer to malloc()'d BIOMOLECULE linked list

Reads the biomolecule assembly data from REMARK 300 and REMARK 350 headers.

The returned pointer is a linked list of BIOMOLECULE structures. The first item in the list contains the information from REMARK 300: the total number of biomolecule assemblies and any additional information provided in REMARK 300 beyond the standard comments. The additional information (biomolecule->details) is a STRINGLIST linked list.

The list of chains will appear in biomolcule->chains as a comma separated list and the transformation to be applied to them will appear in the biomolecule->biomt structure. This is also a linked list so if multiple transformations are required these will appear in further entries to this list.

Additional entries will be found in the BIOMOLECULE linked list for each different biomolecule.

See the test code in PDBHeaderInfo.c for example usage.

26.06.15 Original By: ACRM

Definition at line 1232 of file PDBHeaderInfo.c.

BOOL blGetCompoundWholePDBChain	(	WHOLEPDB *	wpdb,
		char *	chain,
		COMPND *	compnd
	)

Parameters

[in]	*wpdb	WHOLEPDB structure
[in]	*chain	Chain label of interest
[out]	*compnd	Data from the COMPND records

Returns: BOOL Success

Extracts the COMPND data for a specified chain. Returns FALSE if the chain isn't found

28.04.15 Original By: ACRM
04.06.15 Modified to use the ExtractField() routine instead of duplicating code here. Fixes a bug in dealing with compounds where the referenced chains span more than one line

Definition at line 396 of file PDBHeaderInfo.c.

BOOL blGetCompoundWholePDBMolID	(	WHOLEPDB *	wpdb,
		int	molid,
		COMPND *	compnd
	)

Parameters

[in]	*wpdb	WHOLEPDB structure
[in]	molid	MOL_ID from PDB-format COMPND record.
[out]	*compnd	Data from the COMPND records

Returns: BOOL Success

Extracts the COMPND data for a specified MOL_ID. Returns FALSE if the MOL_ID isn't found

13.05.15 Original based on blGetCompoundWholePDBChain(). By: CTP

Definition at line 600 of file PDBHeaderInfo.c.

BOOL blGetHeaderWholePDB	(	WHOLEPDB *	wpdb,
		char *	header,
		int	maxheader,
		char *	date,
		int	maxdate,
		char *	pdbcode,
		int	maxcode
	)

Parameters

[in]	*wpdb	WHOLEPDB structure pointer
[out]	*header	String containing header text
[in]	maxheader	Max length for storing header
[out]	*date	Date string
[in]	maxdate	Max length for storing date
[out]	*pdbcode	PDB code
[in]	maxcode	Max length for storing PDB code

Returns: TRUE: Found HEADER FALSE: Didn't find HEADER

Obtains information from the PDB HEADER record

26.03.15 Original By: ACRM

Definition at line 182 of file PDBHeaderInfo.c.

MODRES* blGetModresWholePDB ( WHOLEPDB * wpdb )

Parameters

[in]

Definition at line 882 of file PDBHeaderInfo.c.

char* blGetSeqresAsStringWholePDB	(	WHOLEPDB *	wpdb,
		char **	chains,
		MODRES *	modres,
		BOOL	doNucleic
	)

Parameters

[in]	*wpdb	Pointer to whole PDB structure
[out]	**chains	Chain labels for the chains - may also be set to NULL if you don't want to record chain labels
[in]	*modres	Linked list of MODRES information. May be NULL if you don't want to translate non-standard amino acids.
[in]	doNucleic	Read sequence for nucleic acid chains

Returns: malloc()'d Sequence from SEQRES, chains separated by a *

Reads sequence from SEQRES records into a character string in 1-letter code. Chains are terminated by * characters.

21.08.97 Original by: ACRM
22.08.97 Added chains parameter
26.08.97 No longer reads DNA/RNA
07.03.07 Added code to check for modified amino acids Now reads from wpdb rather than from the file
07.11.14 Initialize lastchain
11.06.15 Moved to bioplib - doNucleic is now a paramater instead of a global; chains is now an array of strings
12.06.15 Frees memory and returns NULL if no SEQRES found

Definition at line 759 of file PDBHeaderInfo.c.

HASHTABLE* blGetSeqresByChainWholePDB	(	WHOLEPDB *	wpdb,
		MODRES *	modres,
		BOOL	doNucleic
	)

Parameters

[in]	*wpdb	Pointer to whole PDB structure
[in]	*modres	Linked list of MODRES information. May be NULL if you don't want to translate non-standard amino acids.
[in]	doNucleic	Read sequence for nucleic acid chains

Returns: A hash of 1-letter code sequences indexed by chain label

Reads sequence from SEQRES records in 1-letter code, storing the results in a hash indexed by chain label.

25.11.15 Original by: ACRM

Definition at line 1300 of file PDBHeaderInfo.c.

BOOL blGetSpeciesWholePDBChain	(	WHOLEPDB *	wpdb,
		char *	chain,
		PDBSOURCE *	source
	)

Parameters

[in]	*wpdb	WHOLEPDB structure
[in]	*chain	Chain label
[out]	*source	SOURCE information for chain

Returns: Success (chain found?)

Extracts the SOURCE data for a specified chain

26.03.15 Original By: ACRM
13.05.15 Fixes...

Definition at line 534 of file PDBHeaderInfo.c.

BOOL blGetSpeciesWholePDBMolID	(	WHOLEPDB *	wpdb,
		int	molid,
		PDBSOURCE *	source
	)

Parameters

[in]	*wpdb	WHOLEPDB structure
[in]	molid	MOL_ID from PDB-format SOURCE record.
[out]	*source	SOURCE information for chain

Returns: Success (chain found?)

Extracts the SOURCE data for a specified MOL_ID. Returns FALSE if not found.

12.05.15 Original based on blGetSpeciesWholePDBChain(). By: CTP

Definition at line 679 of file PDBHeaderInfo.c.

char* blGetTitleWholePDB ( WHOLEPDB * wpdb )

Parameters

[in] *wpdb WHOLEPDB structure

Returns: Tit;le from PDB file (malloc()'d)

Extracts the title from a PDB file malloc()ing a string in which to store the data. This must be freed by user code

28.04.15 Original By: ACRM
11.05.15 Return NULL if TITLE line absent. By: CTP
09.06.15 Add columns 11 to 80 to title string for both start and continuation lines. By: CTP

Definition at line 228 of file PDBHeaderInfo.c.

Macros

Functions

Detailed Description

Description:

Usage:

Revision History:

Macro Definition Documentation

Function Documentation