search

seung-lab / RealNeuralNetworks.jl

A unified framework for skeletonization, morphological analysis, and connectivity analysis.
Apache License 2.0
39 stars 2 forks source link

Package status? #82

Open wsphillips opened 2 years ago

wsphillips commented 2 years ago

Hi @jingpengw,

What's the status of RealNeuralNetworks.jl at the moment? Is it mainly in maintenance mode? Do you plan to develop it further?

Regardless, I was quite happy to stumble upon this package. I may want to integrate it with some simulator/modeling tooling that I'm working on. Happy to help here and there if it needs some attention, too.

Cheers,

Wiktor

xiuliren commented 2 years ago

Hi @wsphillips , thank you for your interest and willingness to help.

Currently, I am focusing on another project, and this repo is basically in maintenance mode. would you like to share which part of this repo is useful for you? How would you like to use it?

wsphillips commented 2 years ago

Let me start by saying I've only skimmed the source and read through the features--I'm not yet actively using the package, so I might have a bad take on certain details. Nonetheless, I think there's a number of useful things here.

Part of my design goals include strong support for PDE/compartment-based neuron modeling. I want to span a spectrum of detail/biological accuracy. For example, point neurons (like Brian/NEST) --> 1D branched models (like NEURON) --> 3D detailed meshes (like STEPS), etc.

At the most basic level, I would have wound up rolling my own SWC file handling. That alone saves time with importing SWC files from 3D reconstruction repositories (e.g. Allen + neuromorpho.org) Morphological feature extraction and analysis is also super cool in the context of simulations, too (I've been interested by efforts like the TREES toolbox and MorphoPy).

And in general, I want to minimize friction interfacing real-world data. An on-ramp for cryoEM data would be a welcome step in that direction.

xiuliren commented 2 years ago

this repo is designed to interact with real-world data. The paper about it just came out: https://www.frontiersin.org/articles/10.3389/fninf.2022.828169/full

This should be helpful for your understanding. Note that I also have some notebooks to demonstrate how to use it. https://github.com/jingpengw/realneuralnetworks-notebook

wsphillips commented 2 years ago

this repo is designed to interact with real-world data. The paper about it just came out: https://www.frontiersin.org/articles/10.3389/fninf.2022.828169/full

Ah, perfect timing :smile: Thank you for the links to resources!