The second Hackaton was also a big succes.
12 people spent time on presenting what people were working on, after which qith heavy collaboration and discussion people set of to work on their seperate projects while advising others on knowledge they already had. In this way an interacitve afternoon brought various results.
Package Fiji into Docker to be run into BiaFlows. Packaging works, but not in BiaFlows, because it’s 3 years old. To run a Fiji script it needs to be run headless. Can also be run in webpage. Not BiaFlows but we could make something for NL-BioImaging!
Several people compared their analysis scripts, revealing similar results. This is a good sign for reproducability of the data
Cellpose vs Stardist
Compare Cellpose and StartDist nuclei segmentation.
Started to work on a script to measure intensity in protein channel using another foci channel as mask
NKI Foci analyzer to work with OMERO data
Convert NKI Foci Analyzer script to work with OMERO data. That works now in principle; encountered some limitations with possible workaround. In the very end Rodrigo also created this workaround.
Live Foci tracking
Live foci tracking in moving (and rotating) cells, using StarDist, TrackMate, registration of nuclei, and then foci tracking.
Bram & Rolf’s preliminary scripts didn’t work directly on Maarten’s data. -> needs some polishing.
The current registration method (hyperstackreg) is not very robust. Perhaps detect foci in a single nucleus first and then register on the point clouds. We keep in touch.
Tried deconvolution of noisy live foci data. Didn’t work very well, possibly because of high (noisy) background signal inside the cell.
Comparing foci 2D deconvolution with wavelets filtering, on very dim foci. The wavelet filter shows promising results.