Krisztina Boda - UK-QSAR
Download
Report
Transcript Krisztina Boda - UK-QSAR
SYNSPROUT
A tool for generation
synthetically accessible ligands
by De Novo design
IC
AM
S
Krisztina Boda ([email protected])
Supervisor: Prof. Peter Johnson
Topics of Discussion
Sprout Components
De Novo structure generation in Classic SPROUT
SynSPROUT approach
Automatic Fragment Library Generation
Retro-synthetic knowledge base
Fragmenting MDDR
Synthetic knowledge base
Structure generation in SynSPROUT
Example
Future work
SPROUT Components
Detecting potential binding pockets of the protein
structures
Identification of
interaction sites
favourable
hydrogen
bonding
Docking structures at the target interaction sites
Generating 3D molecular structures of novel ligands
by linking the docked starting fragments together
Tools for scoring, sorting and clustering the answer
set
Structure Generation in Classic Sprout
1st Phase
Primary molecular
structure generation
2nd Phase
Conversion of
structure graph into
molecule
CAESA
Generalised
Fuse
fragments
Sequential method
to build structure
graph
Heuristics to avoid
combinatorial
explosion
New
bond
Atom substitution
N
N
O
O
N
Estimating of
synthetic
accessibility
Spiro
O
O
Synthetic Sprout Approach
Build synthetic constraints
into structure generation
Retro
Synthetic
Knowledge
Base
Synthetic
Knowledge
Base
Reliable high
yielding
reactions
Ease of synthesis is a key factor in
drug development
Generating structures and postpruning them is inefficient
BREAKING
STRUCTURES
INTO SMALLER
COMPONENTS
VIRTUAL SYNTHESIS
IN RECEPTOR CAVITY
Readily synthetisable
putative
ligand structures
Drug-like
structures
Fragment
Library
Pool of readily
available
starting
materials
Automatic Fragment Library Generation
2D Drug-like
Structures
Atom & Ring Perception
Retro-Synthetic
Knowledge Base
Detecting RetroSynthetic Patterns
Retro-synthetic patterns
Fragmentation
Retro-synthetic rules
Clustering Identical
Fragments
Synthetic
Knowledge Base
Functional Groups
Synthetic rules
Synthetic Fragment
Library
Perception
Knowledge Base
Aromatic
Normalisation
Hybridisation
H-bonding
properties
Detecting Functional
Groups
Filter
3D Fragment Generation
Corina
Conformation Generation
Omega
Retro-Synthetic Knowledge Base
Retro-Synthetic Patterns
Describing atom
SPCENTER
HETS
HS
EPS
ARYL
CONNECTS
RING
CHEMICAL-LABEL <Amide>
C[HETS=2](=O)–N–C
CHEMICAL-LABEL <Ether>
C[HETS=1]–O-C[SPCENTER=3][HETS=1]
H
N
O
S
O
Describing bond
–, =, #, %
CHAIN
RING
Describing ring
Branching atoms
N
O
O
OH
CHEMICAL-LABEL <Sulphonamide>
O=S(=O)-N
O
O
S
N
H
Retro-Synthetic Knowledge Base
Retro-Synthetic Rules
EXPLANATION Amide Formation
IF Amide
THEN
delete-bond between 1 and 3
add-atom – O to 1
END-THEN
H2 N
S
N
O
O
1
2
O
3
1
3
H
N
O2
4
OH
OH
S
HO
O
N
O
O
OH
OH
Cl
EXPLANATION Ether Formation
IF Ether
THEN
delete-bond between 2 and 3
add-atom – O, Cl, Br to 3
END-THEN
OH
O
OH
Br
O
Fragmenting MDDR
Enamine Formation
Michael Addition
Disulphide
Thioester
Acetal fom Carbonyl
Seconadry Enamine
Acetal from Hemiacetal
Urethane
Aminal from Hemiaminal
Aldol Formation
Urea
Conjugated Enone
Isocyanate from amine/acid
Imine Formation
Friedel Crafts Acylation
Sulphide
Sulphonamide
Ester Formation
Wittig Reaction
Ullman reaction
Alkylation Alpha to Carbonyl
Ether Formation
Amide Formation
Grignard Reaction
Reductive Amination
78515 Drug-like structures
403780 Detected retro-synthetic
patterns
404187 Generated 2D fragments
73141 Unique fragments
Ranking by frequency of
occurrences
1
+
1
0
10000
20000
Filtering out undesirable
fragments
30000
40000
50000
60000
Distribution of applied retro-synthetic rules
Fragment
Library
70000
80000
Synthetic Fragment
Number of occurrences
407
H
H
A
AD
A
H
Carboxylic Acid
(Acceptor)
H
Phenol
(Acceptor-Donor)
Synthetic Knowledge Base
CHEMICAL-LABEL <Carboxylic Acid>
C[SPCENTRE=2];[ARYL=NO](=O)-O[HS=1]
CHEMICAL-LABEL <Primary Amine>
C-N[HS=2];[CONNECTION=1]
Available Func.Groups
Carboxylic Acid
Primary / Secondary Amine
Phenol
Aryl Halide
Carbonyl
Alcohol
Joining Rules
EXPLANATION Amide Formation
IF Carboxylic Acid INTER Primary Amine
THEN destroy-atom 3
form-bond - between 1 and 5
change-hybridization 5 to SP2
DIHEDRAL 0 0
DIHEDRAL 0 180
BOND-LENGTH 1.35
END-THEN
Steps of formation
Hybridization change
Bond type
Bond length
Dihedral penalty/angle
Available Synthetic Rules
2
O
H
+
1
OH
3
O
N5
H
4
N
H
Amide Formation
Ether Formation
Ester Formations
Ullman Reaction
Aldol
Wittig
Reductive Amination
Docking
starting
fragment to
sites
De Novo Drug Design in SynSPROUT
Selecting two
sites to be
connected
Donor site
H2N
Bi-directional
consecutive
search
Overlapping
common
fragments
NH2
O
OH
Acceptor Site
Combined
structures are
re-docked
OH
Docking
starting
fragment to
sites
De Novo Drug Design in SynSPROUT
1.Amide Formation ( Carboxylic Acid –Primary Amine )
Selecting two
sites to be
connected
Donor site
O
O
1
Bi-directional
consecutive
search
Overlapping
common
fragments
O
NH
O
OH
C a rb o xylic
A cid
O
OH
Acceptor Site
NH2
Combined
structures are
re-docked
BF search towards
acceptor site
O
OH
P rim a ry
A m in e
Docking
starting
fragment to
sites
De Novo Drug Design in SynSPROUT
1.Amide Formation ( Carboxylic Acid –Primary Amine )
2.Reductive Amination ( Carbonyl – Primary Amine )
2
Selecting two
sites to be
connected
Bi-directional
consecutive
search
Donor site
N
H
O
OH
OH
P rim a ry
A m in e
Overlapping
common
fragments
H2N
Acceptor Site
Combined
structures are
re-docked
DF search towards
acceptor site
O
O
C a rb o n yl
OH
OH
Docking
starting
fragment to
sites
De Novo Drug Design in SynSPROUT
1.Amide Formation ( Carboxylic Acid –Primary Amine )
2.Reductive Amination ( Carbonyl – Primary Amine )
Selecting two
sites to be
connected
2
N
H
O
O
Overlap
O
OH
O
O
NH
O
Acceptor Site
O
Combined
structures are
re-docked
OH
1 NH
Bi-directional
consecutive
search
Overlapping
common
fragments
Donor site
N
H
OH OH
O
OH
OH
CDK2
Library :
300 fragments/1055 conformations
Docked:935
971
57
6800
Docked:359
Docked:1791
Docked:898
CDK2
Act Score : -7.91
2
3
1
1 Ether formation ( Primary Alkyl Halide – Alcohol )
2 C–S–C ( SH – Primary Alkyl Halide )
3 Amine alkylation ( Primary Amine – Primary Alkyl Halide )
Conclusion
Fragment Library
Generation
De Novo Structure
Generation
Diverse libraries with fragments rich in
biological motifs
Promising structures with estimated high
binding affinity
Future Work
Enhance the program to perform more complex synthetic reactions
Acknowledgements
Peter Johnson
Attilla Ting for atom perception
Colin Martin for building retro-synthetic knowledge base
all members of ICAMS