The HELM Antibody Editor Stefan Klostermann Roche Innovation Center Penzberg Pistoia Alliance Spring Conference, Zürich, 14. April 2015
The HELM Antibody Editor
Stefan KlostermannRoche Innovation Center Penzberg
Pistoia Alliance Spring Conference,
Zürich, 14. April 2015
Agenda
• The motivation for
• The HELM Antibody Editor
• DEMO: From raw sequences to drugs
Classical chemistry
• Aspirin
• Acetylsalicylic acid
SMILES
O=C(Oc1ccccc1C(=O)O)C
InChI
1S/C9H8O4/c1-6(10)13-8-5-3-2-4-7(8)9(11)12/h2-5H,1H3,(H,11,12)
A more biological example
Conjugate of peptide and DNA oligo
• Peptide CHRISTIAN with Ser replaced by dSer
• linked via SMCC (bifunctional linker)
• to oligo CTAG
Classical chemical notationsFull molecular structure - SMILES
• [H]N[C@@H](CSC(=O)C1CCC(CN2C(=O)CC(OC[C@H]3O[C@@H]([C@H](O)[C@@H]3OP(O)(=O)OC[C@H]3O[C@@H]([C@H](O)[C@@H]3OP(O)(=O)OC[C@H]3O[C@@H]([C@H](O)[C@@H]3OP(O)(=O)OC[C@H]3O[C@@H]([C@H](O)[C@@H]3O[H])N3C=NC4=C3N=C(N)NC4=O)N3C=NC4=C3N=CN=C4N)N3C=C(C)C(=O)NC3=O)N3C=CC(N)=NC3=O)C2=O)CC1)C(=O)N[C@@H](CC1=CNC=N1)C(=O)N[C@@H](CCCNC(N)=N)C(=O)N[C@@H]([C@@H](C)CC)C(=O)N[C@H](CO)C(=O)N[C@@H]([C@@H](C)O)C(=O)N[C@@H]([C@@H](C)CC)C(=O)N[C@@H](C)C(=O)N[C@@H](CC(N)=O)C(O)=O
PEPTIDE1{C.H.R.I.[dS].T.I.A.N}|CHEM1{[SMCC]}|RNA1{R(C)P.R(T)P.R(A)P.R(G)}$RNA1,CHEM1,1:R1-1:R2|PEPTIDE1,CHEM1,1:R3-1:R1$$$
HELMNotation and editor views of given example
PEPTIDE1{C.H.R.I.[dS].T.I.A.N}|CHEM1{[SMCC]}|RNA1{R(C)P.R(T)P.R(A)P.R(G)}$RNA1,CHEM1,1:R1-1:R2|PEPTIDE1,CHEM1,1:R3-1:R1$$$
• Limited set of biological building blocks is packed into monomer library
• Notation supports homology searching (a least on natural analogues)
HELMBackground
• Hierarchical Editing Language for Macromolecules– Notation– HELM Editor & Toolkit
• Released by Pfizer to the for public distribution– http://openhelm.org/ Github
• Emerging new notation standard for complex biologicals
Agenda
• The motivation for
• The HELM Antibody Editor
• DEMO: From raw sequences to drugs
12 3
4
HELM for Antibodies
• Why antibodies? – Highly potent and specific drugs especially in Oncology / Immunology
• What makes antibodies a bit special?– Multiple peptide chains – Intra- + inter- chain Cys-Cys bonds– Added functional modules / payloads– Varying complex designs– Designed mutations
LCHC
LCHC
HELM Antibody Editorto handle complex antibodies
• Analysis / decomposition
• Visualization
• Manipulation
• Registration
• Detection, analysis and annotation of all domains
• Cys-Cys bonds– Intra-domain (Ig-like)– Inter-domains (scFv, scFab)– Inter-chains (HC-HC/LC)– Alert if free Cys left
HELM Antibody EditorDecomposition & Assembly
• Variable, hinge, constant
• V: germlines, C: isotypes
• Species ( human?)
• Mutations (knob-into-hole)
• Peptides, linkers, payloads
bispecific
trispecific
effector protein
knob-into-hole
Xmab
HELM Antibody EditorEnable Registration & Data Analysis
Enable registration
• Unique identifier (check)
• Full chemical structure
• Domains & annotations
• Antibody format (used modules)
• Antibody depiction
• Goal: Fully automatic even starting with raw sequences!
Data analysis
1) SAR analysis of
• Specificities
• Antibody formats
• Linker types/lengths
• Payloads (+position)
2) Retrieval of components
21
3
Agenda
• The motivation for
• The HELM Antibody Editor
• DEMO: From raw sequences to drugs
DEMONSTRATIONHELM Antibody Editor
HELMHELM Editor view of demoed complex antibody drug
Role of HELMfor the HELM Antibody Editor
• HELM Notation used for registration
PEPTIDE1{C.H.R.I.[dS].T.I.A.N}|CHEM1{[SMCC]}|RNA1{R(C)P.R(T)P.R(A)P.R(G)}$RNA1,CHEM1,1:R1-1:R2|PEPTIDE1,CHEM1,1:R3-1:R1$$$
• HELM Toolkit used for coding
• HELM Editor used for manipulations of domains (e.g. for ADCs)
Summary
The HELM Antibody Editor
• enables an easy (usually automatic) registration of even complex antibody formats
– for SAR analysis within and across projects– and full search & retrieval of all components.
• has been made public in the tradition of the HELM Editor and is availablenow at https://github.com/PistoiaHELM/HELMAntibodyEditor.
AcknowledgementsTo the developers
Domain editor and framework (Pharma Research and Early Development Informatics, Roche Innovation Center Penzberg, Germany)
• Pandu Raharja (Trainee)
• Stefan Zilch (BridgingIT GmbH, Mannheim)
• Marco Lanig (quattro research GmbH, Martinsried)
• Clemens Wrzodek (Roche pREDi, Penzberg)
Domain recognition (quattro research GmbH, Martinsried, GermanyBernhard Schirm, project lead)
• Anne Mund (Trainee)
• Marco Lanig
Doing now what patients need next