Literature review writing

[rmattson1008]

Just a checklist of lit. review topics to check off. The easiest thing to work on when motivation is low

• [x] basic structure of mito. al la highschool
• [ ] Mitometer
• [x] MitoMo
• [x] MiNa
• [x] Fogo et al
• [x] previous scipy work?
• [x] Iqbal et al
• [ ] find the most comparable deep video modeling that is not mitochondria

And more:

• [ ] paragraph motivating the turn to deep learning

[rmattson1008]

probably leave mitoloc out

[rmattson1008]

I think I need to address methods that prioritize segmentation more (They don't see mito. as one holistic network, apply deep segmentation, analyze pieces of mito.) Example: Fischer, C. A. et al. MitoSegNet: easy-to-use deep learning segmentation for analyzing mitochondrial morphology. iScience 23, 101601. 2020

[rmattson1008]

Shapes w bio names. perilamellar mitochondria. Is there a communication gap between the bio people and the classic ml people? Is mitochondrion a useful concept or is it just fun to write. Am I leaning too hard into the network perspective? example

[magsol]

Is there a communication gap between the bio people and the classic ml people?

Yes.

Is mitochondrion a useful concept or is it just fun to write?

Can't it be both? :)

Am I leaning too hard into the network perspective?

Depends. I think even more than the network perspective, we leaned too hard into the Gaussian-ness of mitochondria; they're decidedly not Gaussian. That said, I do think it's always worthwhile, when we're stuck, to simplify the problem--maybe take a page from that paper and look at three distinct morphologies, instead of arbitrary graphs--and go with that until we get some results that match our intuition. Then add in more complexity.