“found in translation” –neural machine translation ... “found in...

Click here to load reader

Post on 26-Jun-2020

0 views

Category:

Documents

0 download

Embed Size (px)

TRANSCRIPT

  • “Found in Translation” – Neural machine translation models for chemical reaction prediction

    Philippe Schwaller (@phisch124)

    Theophile Gaudin, Riccardo Pisoni, David Lanyi, Costas Bekas & Teodoro Laino

    IBM Research Zurich, Switzerland

    Alpha Lee University of Cambridge

    Chem. Sci., 2018, 9, 6091-6098

  • Exploring the nearly endless chemical space

    O

    O

    O HO

    N

    S

    O

    NH

    O O

    OH

    O

    O- Na+

    Cl

    Cl

    ClCl

    Cl

    N

    OH O

    OH

    HO

    S

    O

    O

    OH

  • Design Make

    Test

    De Novo Molecules VAE / GAN / RL

    Synthetic route planning Reaction outcome prediction

    Experimental verification Credits to Marwin Segler

  • Chemical reaction prediction

  • Data

    US patents

    Lowe (2012, 2017)

    text-mining

    Cl:1 C:2

    C:3

    C:4 C:8

    N:5 N:6

    C:7 OH:14

    N+:15 O-:16

    O:17

    S:9 O:10

    O:11

    OH:12

    OH:13

    Cl:1

    C:2 C:3

    N+:15

    O:14

    O-:16

    C:4

    C:8

    N:5

    N:6 C:7+ +

    Jin et. al. (NIPS 2017)

    filtering

    SMARTS - reactants>>products - partly correct atom-mapping - >1M reactions

    Benchmark dataset (USPTO_500k)

  • Representing molecules for ML Molecular fingerprints 000010000….0100

    Text-based representations - SMILES / INCHI - CN1C=NC2=C1C(=O)N(C(=O)N2C)C

    Graph-based

    1D 2D 3D

    3D structure

    N

    N

    O

    N

    O

    N

    N

    N

    O

    N

    O

    N

  • Atoms as letters, molecules as words

    SMILES to SMILES prediction with sequence-2-sequence models Schwaller et. al. Chem. Sci., 2018, 9, 6091-6098 / Nam & Kim arXiv:1612.09529

    https://arxiv.org/abs/1612.09529

  • How do Seq-2-Seq models work?

  • Step-by-step

    Inspired by Hendrik’s talk in Zurich

  • Step-by-step with attention

    Link to reaction prediction? - Reactants to products

    Br . C C = C 1 C … Source

    C

    Prediction

  • Attention weights

    Plotting the attention weights at every decoder time step.

  • Attention weights Chloro Buchwald-Hartwig amination: US20170240534A1

    Source: Reactants

    Ta rg

    et : P

    ro du

    ct

  • SMILES-2-SMILES

    Trained end-2-end Template-free

    Fully data-driven

    No chemical knowledge incorporated

    No atom-mapping required

    Attention weights and confidence score

    Less than 0.5% invalid SMILES

  • Limitations

    Only as good as the training data

    Difficult to include negative data

    Chemically unreasonable predictions possible

  • What other template-free approaches exist? Graph neural networks • Jin et al. (MIT, 2017): Weisfeiler-Lehman Networks (WLDN) • Network 1: Reaction center identification • Network 2: Product candidate ranking • Trained separately • Outperforms template-based by 10% on USPTO_500k dataset

    • Bradshaw et al. (Cambridge, 2018): Electron path prediction • Gated Graph Neural Networks • Outperforms WLDN on USPTO_350k dataset

    (subset without more difficult reactions, e. g. cycloadditions) Fundamental limitation: require atom-mapped training sets

  • Data set Jin et. al., Schwaller et. al. Our new model MIT_500k 74.0 % reagents Products

    Reactants & reagents mixed Products

    Data set Jin et. al., WLDN

    Schwaller et. al., Seq-2seq

    Bradshaw et al., GGNN

    MIT_500k 79.6 % 80.3 %

    Top-1 accuracy:

    On USPTO_MIT benchmark dataset. Schwaller et al.: Chem. Sci., 2018, 9, 6091-6098 Jin et al.: NIPS, 2017, 30, 2607-2616 Bradshaw et al.: arXiv:1805.10970

    -subset 350k 84.0 % 87.0 %

    Our new model 88.8 %

    https://arxiv.org/abs/1805.10970

  • What else can we do? Reaction scoring

    CC=C1CCCCC1.Cl>>CCC1(Cl)CCCCC1

    CC=C1CCCCC1.Cl>>CC(Cl)C1CCCCC1

    Score: 0.99

    When we provide the model with reactants>>products

    Score: 0.001

    Mark ovni

    kov

    Anti-Markovnikov

  • IBM RXN for Chemistry Freely available now: research.ibm.com/ai4chemistry

    #RXNFORCHEMISTRY

    Booth #530

  • [email protected] / @phisch124

    IBM RXN

    Chem istry

    for

    researc h.ib

    m.c om

    /ai 4c

    he m

    ist ry

    Questions?

    Philippe Schwaller

    mailto:[email protected]