Skip to content

Latest commit

 

History

History
41 lines (35 loc) · 5.84 KB

README.md

File metadata and controls

41 lines (35 loc) · 5.84 KB

Synaptome architecture shapes regional dynamics in the mouse brain

This repository contains code and data in support of "Synaptome architecture shapes regional dynamics in the mouse brain", available as a preprint on bioRxiv. Most of the code was written in Python 3.8.10, with some of the gene ontology analyses being done in Matlab R2022a. Below, I describe all the folders and files in detail.

code

The code folder contains all the code used to run the analyses and generate the figures. A description of each file follows (in an order that complements the manuscript):

  • scpt_remap_synaptome.py will put the synaptome data into the various parcellation schemes I use (e.g. 88 bilateral regions for comparing with fMRI data, 137 right hemisphere regions for comparing with tract-tracing data, ...). This file will also save out the synapse type densities I use in subsequent scripts and analyses.
  • scpt_plot_mouse.py is a general script handy for plotting mouse brains (and this is the script where I plot much of Figure 1). Note that the package I use to plot mouse brains (brainglobe-heatmap) requires Python > 3.9, so I had a separate environment where my Python version was 3.9.19 for all the brain plotting. Yes it was annoying, yes there were reasons I stuck to v3.8.10 in my other environment.
  • scpt_hctsa.py is the main script for comparing synapse type densities with hctsa features. Note hctsa is a Matlab toolbox so Matlab was used to compute all the features (this is work that was done by Andrea Luppi). This is also the script where I plot Figure 2.
  • scpt_hctsa_supplement.py is a file where I really dig into the time-series feature list from hctsa for the sake of understanding (and interpreting) the findings. I also do some supplement-type work here, like checking the effects of SNR, motion, and cell type densities. This script produces Figures S2, S3, S7, S8, and S9.
  • scpt_sc.py is where I compare synapse type densities to a weighted and directed structural connectome. This script corresponds to bits of Figure 3.
  • scpt_fc.py is where I compare synapse type densities to a functional connectome, also corresponding to bits of Figure 3.
  • scpt_scfc.py is where I finally combine SC, FC, and the synaptome, into a structure-function type analysis where I also pull in fMRI data from anaesthetized mice. This script corresponds to Figure 4.
  • scpt_gexp.py is where I compare synapse type density with gene expression profiles from the Allen Mouse Brain Atlas. It's a supplementary analysis but ended up getting a main text figure for itself (Figure 5). Gene Ontology was done using files from this Zenodo repository.

data

The data folder contains data files used for the analyses. If you use this data in your own analyses, please cite the associated papers (and even let them know!).

  • All the "mapping" files are used for mapping synaptome regions to SC/FC regions.
  • cellatlas_ero2018.csv has cell density data for 9 different cell types (some of the "types" are combinations of other types, e.g. "neurons" which is "inhibitory neurons" and "excitatory neurons" together). This data is from Ero et al 2018 Front Neuroinformatics (Data Sheet 2).
  • synaptome contains synapse type densities in their raw form (Type_density_Ricky.xlsx) and mapped to the three parcellations I use (synapse density data was shared by Zhen (Ricky) Liu and Seth Grant, see this paper). This folder also contains synapse protein lifetimes, which were derived by Bulovaite et al.
  • function contains the BOLD time-series for each mouse in the awake state. This also contains the SNR map, and the hctsa outputs for each mouse. This data was shared by Alessandro Gozzi (see this paper).
  • structure contains the structural connectome file which is originally from the supplement of Oh et al 2014.
  • gene_expression contains outputs from abagen, including gene expression sampled from coronal slices, sagittal slices, plus some gene information like entrezID and structure_info.

results

The results folder contains some saved outputs from my scripts, including the some spreadsheets containing the hctsa correlations. The confusing naming convention in results/HCTSA/ is:

  • hctsa- because there are correlation results between synapse type densities and hctsa features
  • norm- because hctsa features were all normalized
  • zscored- because fMRI time-series were z-scored prior to running hctsa on them
  • noexcl- because there was an earlier version of this analysis where I was missing one mouse; "no exclusions" means I'm not excluding any of the mice
  • hits- shows up in the .xlsx spreadsheets and corrs_ shows up in the .npz files; they mean the same thing
  • p-bonferroni-corrected is in the spreadsheets because I include a column with Bonferroni-corrected p-values
  • Awake because the mice are in the awake state

manuscript

The manuscript folder contains the PDF of the manuscript as well as Supplementary Table S1 (hctsa correlations).