Formatting Biological Big Data to Enable (Personalized) Systems Pharmacology

on YouTube ↗36m

" summary

A publicly released resource called Chemical Checker turns experimental bioactivity for ~1,000,000 small molecules into 25 complementary 128‑dimensional vector spaces and uses embedding (Node2Vec) plus siamese networks to infer missing signatures; a separate knowledge‑graph workflow yields 1,000 biological descriptors per gene/cell/disease via random walks and NLP. Merging both spaces enables matching/reversion searches that yield experimentally validated hits — e.g. compounds that reverse Alzheimer transcriptional modules and rescue behavior in mice (4/6) and 14/19 compounds that phenocopy IL‑2R knockdown by blocking IL‑2–dependent proliferation. Raw primary assays are not redistributed; signatures and code are available.

AI-written summary · about

" chapters (21)

1. 0:00Formatting biological Big Data to enable systems pharmacology
2. 0:35coming a long-lasting personal (AI) frustration
3. 2:04Formatting Chemistry
4. 3:21Chemical fingerprints to catalogue small molecules - Proteins
5. 3:54The similarity principle - Molecules with similar structures have similar bioactivities
6. 4:43The Chemical Checker (CC)
7. 6:04Big numbers are powerful, but difficult to interpret
8. 9:57Chemical and biological fingerprints
9. 11:36Large-scale bioactivity annotation Compound
10. 12:26Mapping uncharacterised compounds collections
11. 13:27Signature-Activity Relationship (SigAR) models
12. 14:06Formatting Biology
13. 15:10Harmonizing biological Big Data in a Knowledge Graph
14. 15:54Context-depended biological descriptors
15. 16:59Blending chemistry and biology
16. 17:09Towards a General Connectivity Map (bio-chem)
17. 17:35Signature reversion in Alzheimer's disease (AD)
18. 20:22Global strategy to fight AD in vivo
19. 23:31Mimicking the activity of biodrugs (IL-2R)
20. 26:37Biological signatures, the elephant and the blind men
21. 28:21Structural Bioinformatics & Network Biology group

" related talks

• 30mLeveraging Molecular Data for Drug DiscoveryHaiyan Huang · Computational Challenges in Very Large-Scale 'Omics'
• 53mThe Algorithm for Precision MedicineMatthew Might · DeclMed 2023
• 30mMachine Learning: A New Approach to Drug DiscoveryDaphne Koller · Women in Data Science (WiDS) Stanford 2020
• 10mGenesis Final Round: Nanyang BiologicsDuy Trieu · SuperAI Singapore 2025
• 20mTranslating a Trillion Pointo Diseases of Data into TherapiesAtul Butte · Data By the Bay 2016
• 43mPanel on Interpretability in the Biological SciencesAnshul Kundaje, Jack Gallant, Debbie Marks · Interpretable Machine Learning in Natural and Social Sciences (2022)
• 44mNetwork Propagation as a Tool for Deciphering Disease MechanismsRoded Sharan · Network Biology 2016
• 50mThe algorithm for precision medicineMatt Might · Crunch Conference 2019