Target Generation

This module is used to define PMHC objects (including Target and Template objects) and Model objects. The class used to generate a PANDORA target is PANDORA.PMHC.Target.

This module also contain the Model class, carrying various information of the output models. These objects can be saved when using PANDORA.Pandora.Pandora.model with the argument pickle_out=True

Basic target Example:

>>> from PANDORA.PMHC import Target
>>>
>>> target = PMHC.Target(id = 'myTestCase',
>>>     MHC_class = 'I',
>>>     allele_type = 'HLA-A*02:01',
>>>     peptide = 'LLFGYPVYV',
>>>     anchors = [2,9])
>>>

PMHC

class PANDORA.PMHC.PMHC.PMHC(id, peptide='', allele_type=[], MHC_class='I', M_chain_seq='', B2M_seq='', N_chain_seq='', anchors=[], helix=False, sheet=False)

Bases: ABC

pMHC class. Acts as a parent class to Template and Target

Parameters:

• id – (string) PDB identifier

• allele_type – (list) list of MHC alleles (or allele)

• peptide – (string) peptide sequence

• MHC_class – (string) either ‘I’ or ‘II’ denoting MHC class I and MHC class II respectively

• M_chain_seq – (string) M chain sequence for the Alpha chain

• N_chain_seq – (string) N chain sequence for the Beta chain

• anchors – (list) list of integers specifying which residue(s) of the peptide should be fixed as an anchor during the modelling. MHC class I typically has 2 anchors, while MHC class II typically has 4.

class PANDORA.PMHC.PMHC.Template(id, peptide='', allele_type=[], MHC_class='I', M_chain_seq='', B2M_seq='', N_chain_seq='', anchors=[], G_domain_span=False, helix=False, sheet=False, pdb_path=False, pdb=False, remove_biopython_object=False, reverse=False)

Bases: PMHC

Template structure class. This class holds all information of a template structure that is used for

homology modelling. This class needs a id, allele and the path to a pdb file to work. (sequence info of the chains and peptide can be fetched from the pdb)

Parameters:

• id – (string) PDB identifier

• allele_type – (list) list of MHC alleles (or allele)

• peptide – (string) peptide sequence

• MHC_class – (string) either ‘I’ or ‘II’ denoting MHC class I and MHC class II respectively

• M_chain_seq – (string) M chain sequence for the Alpha chain

• N_chain_seq – (string) N chain sequence for the Beta chain

• anchors – (list) list of integers specifying which residue(s) of the peptide should be fixed as an anchor during the modelling. MHC class I typically has 2 anchors, while MHC class II typically has 4.

• G_domain_span (list) – span of the G domain(s) over the sequence. The format should be [(1, 90),(1, 86)]

• pdb_path – (string) path to pdb file. The user should provide this argument if they want to use a template outside the normal templates folder.

• pdb – (Bio.PDB) Biopython PBD object

get_pdb_path()

parse_pdb(custom_map={'MSE': 'M'})

Loads pdb from path, updates self.pdb field and self.chain_seq/self.peptide if they were empty

Parameters:

custom_map (dict) – custom map of 3-letter to 1-letter residues translation, used by Bio.SeqUtiles.seq1 to decide how to assign non-canonical residues. Defaults to {“MSE”:”M”}.

info()

Print the basic info of this structure

calc_contacts()

calc_anchors()

calc_anchor_contacts()

check_allele_name()

Checks the spell of the allele name and tried to correct it. Prints out warning if the allele name does not seem correct.

class PANDORA.PMHC.PMHC.Target(id, peptide, allele_type=[], MHC_class='I', M_chain_seq='', N_chain_seq='', B2M_seq='', anchors=[], helix=False, sheet=False, templates=False, use_netmhcpan=False, use_templ_seq=False, output_dir=False, rm_netmhcpan_output=True, reverse=False)

Bases: PMHC

Target structure class. This class needs an ID (preferably a PDB ID), allele and pepide information.

Parameters:

• id – (string) PDB identifier

• peptide – (string) peptide sequence

• allele_type – (list) list of MHC alleles (or allele)

• MHC_class – (string) either ‘I’ or ‘II’ denoting MHC class I and MHC class II respectively

• M_chain_seq – (string) M chain sequence for the Alpha chain

• N_chain_seq – (string) N chain sequence for the Beta chain

• anchors – (list) list of integers specifying which residue(s) of the peptide should be fixed as an anchor during the modelling. MHC class I typically has 2 anchors, while MHC class II typically has 4.

• templates – Template object. The user can specify that PANDORA uses a certain structure as template.

• use_netmhcpan (bool) – If True, uses local installation of NetMHCPan to predict the anchors when anchor positions are not provided. Defaults to False.

• use_templ_seq (bool) – If True, if no MHC chain sequences could be retrieved starting from the allele name, it will use the best template MHC sequences for the modelling.

• output_dir – (string) Path to output directory. Defaults to current working directory.

• rm_netmhcpan_output – (bool) If True, removes the netmhcpan infile and outfile after having used them for netmhcpan.

info()

Print the basic info of this structure

calc_contacts()

calc_anchor_contacts()

check_allele_name()

Checks the spell of the allele name and tried to correct it. Prints out warning if the allele name does not seem correct.

retrieve_MHC_refseq(input_file=None, chain='M', permissive=False)

Retrieves MHC reference sequence from fasta file.

Parameters:

• input_file (str, optional) – Path to the input reference fasta file. Defaults to None.

• chain (str) – ID of the chain to be retrieved (M or N). Defaults to ‘M’.

• permissive (bool) – If True, if no chain of the exact allele can be found, it will try to retrieve a chain from the same allele subgroup. Defaults to False.

Returns:

None.

fill_allele_seq_info(use_templ_seq=False)

Fills in MHC-II alpha chain name if missing and it tries to retireve the sequence according to the allele name of vice versa

Parameters:

use_templ_seq (bool, optional) – If true, it uses the template MHC sequence for each chain a sequence could not be found. This function is mainly for benchmarking reason. Defaults to False.

Raises:

• Exception – _description_

• Exception – _description_

• Exception – _description_

make_output_dir()

Create an output directory and move the template pdb there

Uses self.output_dir (str): Path to output directory. Defaults to os.getcwd().

Parameters:

None

Model

class PANDORA.PMHC.Model.Model(target, model_path='', output_dir=False, pdb=False, molpdf=0, dope=0)

Initiate model object :param target: Target object :param model_path: (string) path to hypothetical model :param output_dir: (string) output directory :param pdb: Bio.PDB object of the hypothetical model :param molpdf: (float) molpdf score :param dope: (float) DOPE score

calc_LRMSD(reference_pdb, atoms=['C', 'CA', 'N', 'O'], ligand_zone='whole')

Calculate the L-RMSD between the decoy and reference structure (ground truth).

This function requires the pdb2sql module for L-RMSD calculation.

Parameters:

• reference_pdb – Bio.PDB object or path to pdb file

• atoms (list, optional) – The list of atoms of the ligand selected to calculte the LRMSD . Defaults to [‘C’, ‘CA’, ‘N’, ‘O’].

• ligand_zone (str, optional) – The region of the Ligand selected to calculate the LRMSD. Defaults to “whole”.

Raises:

Exception – PDB2SQL LRMSD claulation failed

Returns: (float) L-RMSD calculated by PDB2SQL

PANDORA.PMHC.Model.merge_chains(pdb)

Merges two chains of MHCII to one chain. pdb2sql can only calculate L-rmsd with one chain. :param pdb: Bio.PDB object

Returns: Bio.PDB object with its M and N chain merged as M chain

PANDORA.PMHC.Model.renumber(pdb_ref, pdb_decoy, custom_map={'MSE': 'M'})

aligns two pdb’s and renumber them accordingly. :param pdb_ref: Bio.PDB object :param pdb_decoy: Bio.PDB object

Returns: Bio.PDB objects with renumbered residues

PANDORA.PMHC.Model.homogenize_pdbs(decoy, ref, atoms, output_dir, target_id='MHC', anchors=False, flanking=False)

Make sure that the decoy and reference structure have the same structure sequences. :param decoy: Bio.PDB object of the decoy structure :param ref: Bio.PDB object of the reference structure :param output_dir: (string) directory that is used to write intermediate files

Returns: (tuple) Bio.PDB objects with the same structure sequence

PANDORA.PMHC.Model.get_Gdomain_lzone(ref_pdb, output_dir, MHC_class)

Produce a lzone file for pdb2sql. :param ref_pdb: path to the pdb file to use for the lzone :type ref_pdb: str :param output_dir: output directory :type output_dir: str :param MHC_class: Class of the MHC :type MHC_class: str

Raises:

Exception – In case there are unexpected chain names it raises an exception

Returns:

Path to the output file

Return type:

outfile (str)

PANDORA.PMHC.Model.remove_C_like_domain(pdb, need_to_be_removed=None)

Removes the C-like domain from a MHC struture and keeps only the G domain :param pdb: (Bio.PDB): Bio.PDB object with chains names M (N for MHCII) and P :param need_to_be_removed: list of atoms to remove from M chain. Defaults to None. :type need_to_be_removed: list, optional

Returns: (Bio.PDB): Bio.PDB object without the C-like domain

PANDORA.PMHC.Model.trim_indels(pdb_ref, pdb_decoy, ref_sequences, decoy_sequences)

Trim indels for both reference and decoy PDBs

Parameters:

• pdb_ref (Bio.PDB) – object with chains names M (N for MHCII) and P

• pdb_decoy (Bio.PDB) – object with chains names M (N for MHCII) and P

• ref_sequences – List of residue numbers in reference PDB

• decoy_sequences – List of residue numbers in decoy PDB

Returns: Bio.PDB objects (reference and decoy) with matched residues

PANDORA.PMHC.Model.get_residue_atoms(decoy_res, ref_res)

Retrieves all atom names of a given residue for both reference and decoy

Parameters:

• ref_res (Bio.PDB.Residue) – Residue object from reference PDB

• decoy_res (Bio.PDB.Residue) – Residue object from decoy PDB

Returns: Bio.PDB.Atom objects (reference and decoy) of the given residues

PANDORA.PMHC.Model.remove_atoms_from_res(ref_chain, decoy_chain, i)

Mismatched residues in Reference is removed both from decoy and reference

PANDORA.PMHC.Model.remove_mismatched_atoms_from_res(diff_atoms, dif_res, atoms)

Mismatched atoms in the given residue object is removed

PANDORA.PMHC.Model.remove_mismatched_atoms_from_pdb(ref, decoy, atoms)

Mismatched atoms in the given PDB object is removed :param ref: object with chains names M (N for MHCII) and P :type ref: Bio.PDB :param decoy: object with chains names M (N for MHCII) and P :type decoy: Bio.PDB :param atoms: _description_ :type atoms: _type_

Returns:

_description_

Return type:

_type_

Anchors

PANDORA.PMHC.Anchors.pMHCI_anchors(pdb)

Finds peptide anchor residues of p:MHCI complexes

Parameters:

pdb – Bio.PDB object (or path to pdb file (string))

Returns: (list of ints) Peptide anchor residue positions

PANDORA.PMHC.Anchors.pMHCII_anchors(pdb)

Finds peptide anchor residues of p:MHCII complexes

Parameters:

pdb – Bio.PDB object (or path to pdb file (string))

Returns: (list of ints) Peptide anchor residue positions