Another year, another BioIT is over. It felt slightly strange to be having it in mid April compared to later May for the past few years. The first day even coincided with Boston marathon but our FAIR workshop coincided with that so could not watch. At least they had a reasonable weather day, not too hot and dry. In fact the weather all week was pleasant except for one wet day, which felt just the UK weather we have been having.
In writing some notes from BioIT 2024, not sure about others but find it harder to find time to go through the conference talks when it’s the more casual interactions that are interesting and productive.
Curlew Research BioIT 2024 Review
GenAI
Multiple tracks – too many to keep track of!!!
Themes across all of the talks
LLMs in general:
-
Trust – verification and validation
-
Stopping hallucinations – see RAG approach
-
Keeping private data private – risk with a unconstrained LLM
Data Management
FAIR talks – see the workshop session below
Snowflake & other approaches and life science panel –
Pharma Informatics
-
J&J Platform updates to cope multi modality and especially larger molecules
-
Agnostic platform created
-
-
Roche – Requesting tools / molecule Cascades (assay screening)
-
Roche – cPanda prediction platform
-
Takeda – platform roadmap – assay, warehouse
FAIR Workshop (Mon 15 Apr) – “FAIR Train”
This was a one day workshop featuring 11 speakers and panels organised by Ishwar Chandramouliswaran @NIH
FAIR journey is in progress for both Life science and wider academics
-
Need to decide on the business questions and value as part of FAIR plans and projects
-
When is “FAIR enough” is something many are reviewing. Often cannot often afford full retrospective FAIRification and the value return is hard to justify
-
What can be done as part of data generation to be FAIR from start
-
FAIR as much about change management as technology
Coupled with exhibition, events and informal discussions this was a good few days, plus the chance to visit folks in Cambridge, Mass