Coffee with EMBL #16: DeepMind and the Future of AI in Life Sciences (Part 1)

This installment of Coffee with EMBL revolved around the breakthrough of DeepMind's AlphaFold in protein folding research, with speakers discussing its potential impact on the academic community and the limitations of current methods. They also discussed recent breakthroughs in AI-driven protein structure prediction and its potential applications. The speakers emphasized the importance of databases and machine learning in structural biology, and highlighted the potential of AI to revolutionize the field. They also discussed the need for understanding AI's limitations and the potential for AI to generate new biomolecules and understand health data.
_ _ _
Guest speakers:
🎙️ Dame Janet Thornton, Director Emeritus (2001-2015), Group leader and Senior Scientist, EMBL-EBI, 1994-2023
Dame Commander of the Order of the British Empire (2000)
Joint Vice President of ERC, 2019-2020
🎙️ Sameer Velankar, Team Leader and Senior Scientist at EMBL-EBI since 2000
🎙️ Chris Sander, Professor in Residence of Cell Biology, Harvard Medical School
Group Leader and Senior Scientist, EMBL Heidelberg, 1986-1998
🎙️ Andrew Miller, First Head of EMBL Grenoble, Group leader, 1975-1980
🎙️ Anna Kreshuk, Group leader at EMBL Heidelberg since 2018
_ _ _
Timestamps:
0:05 - AI's Impact on Protein Folding with EMBL and DeepMind Experts
Angus Lamond welcomes attendees and begins the discussion on AI's impact on research and society.
6:25 - Predicting Protein Folding Using AI
Thomas Vaccari explains protein folding and the role of amino acids, highlighting an AI-based method for accurate predictions.
12:07 - AI-Predicted Protein Structures and Their Accessibility for Research
Samia Velankar discusses the AlphaFold database, its extensive protein structure predictions, and its open accessibility.
17:55 - DeepMind's AI-Powered Protein Structure Prediction and Its Impact
Discussion on making AI research openly accessible and the projected growth of the AlphaFold database.
24:03 - AI-Driven Protein Structure Prediction and Its Limitations
Janet Thornton and Chris Sander discuss the limitations and accuracy issues of AI models in protein structure prediction.
32:42 - Protein Structure Prediction Using AlphaFold and RosettaFold
Chris Sander highlights the caveats of AI predictions, including the limitations in drug design potential.
42:09 - AI in Structural Biology, Including Database and Machine Learning
Angus Lamond and Chris Sander discuss the importance of databases and community contributions in AI-driven structural biology.
47:55 - The Future of Structural Biology Infrastructure and Technology Investment
Andrew Miller emphasizes the importance of investing in structural biology infrastructure, like cryo-EM and synchrotrons.
54:56 - Applying AI to Image Analysis in Biology
Anna Kresha explores AI applications in imaging research and the potential for AI in segmentation, tracking, and behavior analysis.
59:54 - Using AI to Analyze Biological Data and Predict Cellular Interactions
Discussion on building predictive models of cell biology through AI and machine learning.
1:05:56 - AI's Role in Structural Biology, Limitations, and Future Prospects
Discussion on the necessity of human validation for AI predictions and the experimental validation needed for structural biology.
1:10:47 - AI's Potential in Protein Structure Design and Material Development
Chris Sander discusses AI-assisted protein design and its implications for biotherapeutics and material science.
1:15:07 - Predicting Protein Structures and Ligand Interactions Using AI
Challenges of predicting small molecule ligand interactions and the competitive dynamics affecting protein structure prediction.
1:19:52 - AI Applications in Biology: Protein Folding, Health Data Analysis, and Imaging
Discussion on AI's potential in health data analysis, disease prediction, and image annotation in microscopy.
1:24:03 - Using AI in Biology: Cancer Research and Data Sharing
Chris Sander highlights the challenges and potential of AI in clinical research, emphasizing the need for well-organized databases.
1:29:30 - AI Applications in Protein Folding, Image Analysis, and Clinical Research
Angus Lamond and Thomas Vaccari discuss AI's broad impact on society and thank contributors for their insights.
_ _ _
Coffee with EMBL is hosted by EMBL and moderated by EMBL alumnus Angus Lamond, Professor of Biochemistry, University of Dundee | Former Group Leader at EMBL Heidelberg, 1987-1995, who co-organises the series together with Thomas Vaccari, Associate Professor, University of Milan | Former Predoctoral Fellow (Ephrussi Group) at EMBL Heidelberg, 1999-2004) and the EMBL Alumni Relations team.
Learn more about the series at www.embl.org/about/info/alumn...
_ _ _
Access live chat highlights (+ useful links and notes) at oc.embl.de/index.php/s/yIJ0ox...
_ _ _
Recorded on Friday, 17 September 2021