Startup meeting AI/ML/OI/SB tech group 220608

(original location link)

Agenda:

Zoom Link: https://lu-se.zoom.us/j/65446548556

Time: 1100-1200 

Participants: Ashfaq Ali, Bengt Sennblad, Carolina Wählby, Claudio Mirabello, Lars Eklund, Fredrik Nysjö, Daniel Malmodin, Eva Freyhult,  Nima Rafati, Sergiu Netotea, Per Unneberg, Thomas Svensson, Mun-Gwan Hong, Julie Lorent, Erik Fasterius, Paul Pyl, Nikolay, Jonas Söderberg, Olga Dethlefsen

One or two groups? Scope:

• Topics not overlapping?
• Integrative NBIS projects not ML-fit
• Too few people to motivate two groups?
• Initially tech groups will be internal to NBIS
• Associated JCs could be more inclusive (?)
• Nima: how reach out to DDLS then?
• Decision: One group

DDLS conference

• Carolina: We/the techgroup/NBIS should have a showcase for the upcoming DDLS conference (Nov 15-16)

Topics of interest:

• Method-based tech group
• Omics from multiple assays/time points/ experiments
• ML-based
• Network-based
• Knowledge graphs
• Systems biology 
• (== non-ML integrative omics?)
• Imaging and Spatial omics?
• ML approaches (i.e., not only multi omics)?
• ML integration/reuse/generalization  across interests? (Suggested topic LE)
• NN/DL
• Data Engineering and data analyses, reproducibility
• Comment: border against statistical genomics (also method-based) t4ech group, e.g., statistical learning, probabilistic models maybe sorts under this stat genomics tech group.

Format:

• How to meet

• Quarterly or Biannual meeting, with possible involvement of other scilifeLab ? 
• In connection with the retreat (parallel session or similar) 
• If the retreat is very busy, possibly add an extra meeting.
• Should there be a topical workshop/conferences?
• Journal club?
• Positive response from Nikolay
• 3-4/semester or monthly 
• Zoom
• Monthly chat 
• In connection to/before JC
• Strategic event to engage wider audience 
• Impact projects together?
• Should training be linked to the group?

Communication:

• Slack
• Meeting notes
• Use Wabi wiki space
• Use Tech group Calendar
• Monthly chat with possible followup item on the agenda
• Maybe 3rd Friday in the month at 10.00? (or something similar)
• Followed by JC or similar?

Comments:

Overlap between the areas

• Perhaps in Spatial Omics
• Integrative with machine learning maybe too broad?
• Projects NBIS is involved in, maybe sample sizes are too small
• Possible overlap with Spatial Omics, that maybe fine.
• Many examples in Single cell omics (autoencoders)
• Cancer Genomics
• Two independent groups may not practical due to size
• Possible issue with disclosing projects with external stakeholders