netneurolab/hansen_gene-receptor

Data and code in support of Hansen et al., 2022 "Correspondence between gene expression and neurotransmitter receptor and transporter density in the human brain", NeuroImage

Gene expression-receptor density associations

This repository contains code and data created in support of my project, "Correspondence between gene expression and neurotransmitter receptor and transporter density in the human cortex", published in NeuroImage. All code was written in Python.

Main analyses

Code to reproduce figures in the main text can be found in code/main.py. Code for supplementary analyses can be found in code/supplement.py. Running all gene expression-receptor density correlations (PET, autoradiography, microarray, RNAseq, cortex, subcortex, etc...) with spin-tests can be done using code/correlate_expression_density.py.

Data

The data folder contains the following files/directories:

• autoradiography: this contains autoradiography data originally collected by Zilles & Palomero-Gallagher, 2017, and available as numpy files (data originally presented here in support of this paper)
• expression: this contains microarray and RNAseq gene expression data and differential stability for two parcellation resolutions, originally from the AHBA. Data was processed using the abagen toolbox (see this paper). Gene expression was estimated using multiple different probe selection methods (see Supplement hence the repeats.
• PET_receptors.csv: 68 cortical Desikan Killiany regions x 18 neurotransmitter receptor and transporter PET-derived densities, orignally collated and used here for this paper. Note that the order of receptors can be found in main.py
• PET_receptors_scale125.csv: same thing but 219 regions (I'm not the one that came up with this naming convention)
• PET_receptors_subcortex.csv: same thing but for 15 subcortical regions.
• panther_ontologies are lists of genes involved in the protein pathway of each neurotransmitter (e.g. "acetylcholine"), derived from the (PANTHER Classification System)(http://pantherdb.org/panther/globalSearch.jsp?).

Results

The results folder contains:

• AUTcorrs_microarray.csv and AUTcorrs_rnaseq.csv: Spearman $r$ and $p_\text{spin}$ for each autoradiography-derived receptor-gene pair using microarray and RNAseq gene expression data.
• autorad_data.npy: generated region x receptor matrix of autoradiography data, converted from the original 44 regions to the Desikan Killiany parcellation. Receptor names can be found in main.py.
• PETcorrs_scale033_microarray.csv, PETcorrs_scale033_rnaseq.csv, and PETcorrs_scale125_microarray.csv: Spearman $r$ and $p_\text{spin}$ for each PET-derived receptor-gene pair using microarray and RNAseq gene expression data, under two parcellations (Cammoun-033, regionally equivalent to Desikan Killiany, and Cammoun-125, a subdivision of Cammoun-033).
• layercorrs.npz: my saved output of the autoradiography expression-density correlations within three laminar layers.
• panther.csv: the PANTHER Classification output (see Supplement).