EE-graph-dyn : wiring a worm's brain

@ Author: Vito Dichio (email: [email protected] | twitter: @vidichio)

Welcome to the official repository for the code accompanying the paper:

Dichio & De Vico Fallani (2024). Exploration-Exploitation Paradigm for Networked Biological Systems. Physical Review Letters, 132(9), 098402 [1].

This project introduced a novel formalism -- inspired by the evolutionary dynamics -- to model those dynamics in biology that result from an exploration of the configuration space and the exploitation of functional states. On such example, we argued, is the brain wiring problem [5]. Within this context, we used the exploration-exploitation dynamics to model the brain maturation of the nematode C. elegans, starting from the connectomes recently obtained by Witvliet et al. [6]. A detailed explanation of the theoretical and computational methods can be found in the Supplemental Materials of [1] and in the chapters 4-5 of [2].

These scripts are supposed to be a clean, minimal and commented version of the code that has been used to get the results in the paper -- yes, the originals are much messier. You just need to git-clone this repository and run the scripts!

If you would like to re-use this code or build on it, I would be more than glad to help, get in touch!

Brain of a worm

Preliminary steps to the EE dynamics. Coded in R v4.0.4 using RStudio v2022.12.0.

• 1a | Preprocessing. The script witvliet.R takes as input the worm connectomes from [7] (stored in src/data) and outputs the corresponding adjacency matrices (unweighted, undirected) and neuron information.
• 1b | ERG analysis. The script erg.R takes as input one of the two adult brains and its node attributes (stored in src) and performs an ERG estimation based on the model described in [1]. For this purpose, the ergm library is used [4]. See also [3] for a methodological discussion of exponential random graph (ERG) models.

EE for a brain maturation

Exploration-exploitation dynamics for the C. elegans brain maturation. Coded in Python v3.9.7 using Spyder v5.1.5.

The general structure of these simulations mimics the population-based evolutionary simulations described in [6]. See Sec. 4.3 in [2] for a detailed description.

• 2 | EE dynamics The script EE.py takes as input the worm connectome at birth (stored in src/data) and uses it to initialise one EE simulation. The values of the parameters are those corresponding to the optimal simulation, displayed in fig.(2) in [1]. Internal and/or auxiliary functions for this script are coded in src/building_blocks.py and src/erg_terms.py.

Bibliography

1. Dichio & De Vico Fallani (2024). Exploration-Exploitation Paradigm for Networked Biological Systems. Physical Review Letters, 132(9), 098402.
2. Dichio (2023). The exploration-exploitation paradigm: a biophysical approach (PhD Thesis). arXiv preprint arXiv:2312.14850 .
3. Dichio , & De Vico Fallani (2023). Statistical models of complex brain networks: a maximum entropy approach. Reports on progress in physics, 86(10), 102601.
4. Handcock et al. (2008). statnet: Software tools for the representation, visualization, analysis and simulation of network data. Journal of statistical software, 24(1), 1548.
5. Hassan & Hiesinger (2015). Beyond molecular codes: simple rules to wire complex brains. Cell, 163(2), 285-291.
6. Zanini & Neher (2012). FFPopSim: an efficient forward simulation package for the evolution of large populations. Bioinformatics, 28(24), 3332-3333.
7. Witvliet et al. (2021). Connectomes across development reveal principles of brain maturation. Nature, 596(7871), 257-261.