BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:From Code to the Lab: Applying Transfer Learning to Discover New A
 ntibacterials - Sergio Bacallado de Lara\, Department of Pure Maths and Ma
 thematical Statistics
DTSTART:20251027T120000Z
DTEND:20251027T123000Z
UID:TALK237433@talks.cam.ac.uk
CONTACT:Sam Nallaperuma-Herzberg
DESCRIPTION:The scarcity of high-quality public data presents a major hurd
 le for applying deep learning to drug discovery. This is particularly acut
 e in the search for new antibiotics\, where active compounds are rare. In 
 this talk\, I will discuss how my group has addressed this "low-data" chal
 lenge by applying transfer learning with deep graph neural networks (DGNNs
 ) to discover novel antibacterials. By pre-training models on large\, gene
 ral chemical datasets before fine-tuning them on small\, specific antibact
 erial screens\, we were able to virtually screen over a billion compounds.
  Supported by a £15\,000 grant from the Accelerate Programme\, this compu
 tational work led to the experimental validation of several potent\, low-t
 oxicity compounds with broad-spectrum activity against critical ESKAPE pat
 hogens\, achieving a ~54% hit rate from our prioritized candidates.\nThis 
 project highlights a broader story about bridging the gap between computat
 ional theory and experimental practice. I'll touch upon the significant ch
 allenges in developing robust benchmarks for molecular machine learning\, 
 a critical step for ensuring our models are truly learning meaningful chem
 ical principles. Furthermore\, I will share how our group\, based in the m
 aths faculty\, collaborated with experimental facilities to validate our c
 omputational hypotheses\, turning predictive models into tangible results.
  My goal is to encourage others in the Cambridge AI community to develop m
 ethodologies that are not only computationally novel but are also geared t
 owards practical\, hypothesis-driven validation\, demonstrating how even m
 odest funding can generate significant real-world impact.
LOCATION:SS03 Seminar Room\, Willam Gates building (Department of Computer
  Science and Technology)
END:VEVENT
END:VCALENDAR