2024 Q1 and SAC goals

Goals

1. Be able to define baseline variation/randomness, ie null hypothesis (in progress)
2. Determine rules for peroxisomes and endosomes and draft manuscript
3. Apply to variance analysis (2.0), do we have a new way to define sameness between punctate structures?
4. Make cellPACK more usable by people outside our group

Major time commitments from

1. Megan
2. Saurabh
3. Ruge
4. Karthik

Not anticipating devops support being needed until after SAC.

Tasks

New features

• (*1) Combine rules/gradients, with different weights
• (*2) Publish cellPACK on pypi test installation as a third party user.
• Complex weight maps: example using the position of other structures as grid weights
• Long term/after punctate analysis conclusion: Filament packing (either Actin or mitochondria) -->
  • use third party for cylinder packing?
  • have to finish the code for converting to Simularium for visualization

Analysis:

• - (*3) Text debt/maintenance on cellPack analysis repo. Done when we have a pipeline that re-creates the peroxisomes analysis results. In addition to cellPACK make sure other deps are available publicly. Update the README to make other people able to use the pipelines.
  • [Depends on *2]
• (stretch) with new pairwise comparisons, rerun on the variance paper dataset to compare the heat map/clustering result. (for a subset of structures)
  • Use the EMD instead of Pearson correlation between PILRs
• Tune the strength of the nucleus gradient to more accurately reproduce some aspects of the observed data.
  • match the peak of the distribution?
  • minimize the EMD between observed and simulated data (can use rep & learning ML approaches)
  • pack in mean cell an in the 305 real cells for peroxisome data to confirm that cellPACK is behaving

Long term goal of having cellPack run fully in browser:

• (*4) Put cellPack on a server
  • [Depends on *2]
• (stretch) prototype front end site.
• MVP: ability to change a parameter in existing recipe and pack. -->

(5) manuscript of punctate packing

• [Depends on *3]
• Endosome packing
  • [Depends on *1]
• Peroxisomes
• finish up count and size analysis
• request microscopy experiments to support ideas: -->
  • could we get a time-series FOV over 2-3 days and a shorter high-res time series to understand Peroxisome localization cycles for example: do they wander in/out randomly or are they born near the nuclear surface and tend to migrate out slowly and die near the cells surface or are they super-stable and maybe just wander/migrate very slowly out, or even just get born and tend to stay in one area?

--> move to 2025 goals