
0.25
0.5
0.75
1.25
1.5
1.75
2
Deep learning for computational chemistry: compound representation, ADMET profiles and automatic optimization
Published on 2019-06-28110 Views
One of the main challenges in small molecule drug discovery is efficiently finding novel chemical compounds with desirable properties. Such properties can be physico-chemical (like logD or solubility)
Related categories
Presentation
Deep learning for computational chemistry00:00
Introduction00:45
Computational chemistry - 100:47
Computational chemistry - 201:00
Describing the chemical matter02:17
History of deep learning for computational chemistry - 103:37
History of deep learning for computational chemistry - 204:25
Continuous, data-driven molecular descriptors04:52
General idea 05:22
Data and representations07:28
Training the translation model - 108:18
Training the translation model - 209:23
Training the translation model - 309:38
Training the translation model - 409:51
Training the translation model - 510:55
Performance of the autoencoder embedding as molecular descriptor - 112:52
Performance of the autoencoder embedding as molecular descriptor - 213:39
Performance of the autoencoder embedding as molecular descriptor - 314:52
Performance of the autoencoder embedding as molecular descriptor - 414:54
Performance of the autoencoder embedding as molecular descriptor - 516:23
Wrap - up - 116:40
Multitask learning for ADMET prediction17:40
Absorption – Distribution – Metabolism – Excretion – Toxicity19:31
The data20:15
Different methods compared20:58
Multitask learning21:27
Motivation for a multitask approach and expected benefits22:23
Multitask learning in practice - 123:31
Multitask learning in practice - 224:16
Graph convolution operations - 126:19
Graph convolution operations - 228:57
Graph convolution operations - 329:03
Graph convolution operations - 429:31
Going back to a molecule-level representation29:54
Model evaluation30:28
Absorption – Distribution: physico-chemical properties - 131:23
Absorption – Distribution: physico-chemical properties - 232:15
Absorption – Distribution: physico-chemical properties - 333:44
Absorption – Distribution: physico-chemical properties - 434:49
Wrap - up - 237:05
Absorption – Distribution: physico-chemical properties - 538:31
Molecule Swarm Optimization (MSO)42:44
Introduction: navigating the CDDD chemical space42:48
Introduction: navigating the CDDD chemical space - 243:22
How to steer the navigation towards useful chemistry? - 144:29
How to steer the navigation towards useful chemistry? - 246:13
Particle swarm optimization 47:04
Single parameter optimization49:08
Restraining the chemical space - 150:37
Restraining the chemical space - 251:57
Restraining the chemical space - 352:48
Multi-parameter optimization - 154:21
Multi-parameter optimization - 255:51
Wrap - up - 356:34
Conclusion - 157:46
Conclusion - 257:49
Thank you58:21