Page 16 - Read Online

Donskov et al. J Transl Genet Genom 2021;5:136-62 https://dx.doi.org/10.20517/jtgg.2021.12 Page 5

score < 0.8, were excluded from the analyses.

For the analysis of promotor HRE containing genes, available curated and non-redundant sets of
transcription factor binding sites (TFBSs) for NR monomers, HOCOMOCOv11_core_HUMAN, were
[89]
downloaded from the HOCOMOCO collection (http://www.cbrc.kaust.edu.sa/hocomoco11) and
genomic positions were identified using the FIMO tool (http://meme-suite.org/tools/fimo) . Subsequently,
[90]
lists were generated for each NR with genes containing their HRE within their promotor sequences (2000 bp
upstream of TSS). Gene annotation files contained every human protein encoding gene detected in brain
tissue (www.brainspan.org, RPKM ≠ 0 in any sample). For the assessment of similarities between HRE gene
sets, pairwise Jaccard similarity coefficients and significance of overlap were calculated using the
[91]
GeneOverlap R package (version 1.24.0) .

Cortical transcriptomic profiling of NTC and HRE gene sets in patients
We used available analyses of differentially expressed genes (DEGs) in the dorsolateral prefrontal cortex of
258 SZ patients and 271 healthy controls from the CommonMind Consortium (CMC; CommonMind.org
Synapse ID: syn5607652) .
[92]

[93]
Enrichment analysis of TFBSs was carried out according to Gearing et al. using CiiiDER. Briefly,
promotor sequences (2000 bp upstream of TSS) were extracted from the Homo sapiens GRCh38.94 genome
file. Identification of TFBSs in these sequences was performed with HOCOMOCOv11_core_HUMAN
transcription factor position frequency matrices [downloaded from the HOCOMOCO collection
[89]
(http://www.cbrc.kaust.edu.sa/hocomoco11)] and a deficit cut-off of 0.15. CiiiDER enrichment analysis of
overrepresented NR TFBSs in DEG query sequences compared to non-DEG query sequences (from 1000
genes with p~1 and logFC~0) was determined by comparing the number of sequences with predicted TFBSs
to the number of those without, using a Fisher’s exact test.

Brain transcriptomic profile of the nuclear receptor transcriptome complex
Normalized gene expression values (RPKM) for 16 different brain tissues in the developing and mature
brain was downloaded from www.brainspan.org . Developmental stages were defined as Prenatal
[87]
(8-24 pcw); Early childhood (4 months-4 years); Puberty (8-19 years), and Adulthood (21-40 years) and the
average RPKM within groups was plotted with hierarchical clustering (average correlation with row
[94]
centering) using ClustVis . Human brain cell type-specific gene expression annotations were obtained
from McKenzie et al. . The genes displaying sex-biased expression in 16 brain tissues across four
[95]
developmental stages (prenatal, early childhood, puberty and adulthood) were assessed in a publicly
available human dataset (www.brainspan.org) and obtained from Shi et al. . The significance of overlap
[96]
[87]
between NTC gene sets and brain sex-biased genes was analyzed using permutation analysis (n = 10,000
permutations) based on a list of all protein-coding and brain-expressed genes. In each permutation, a gene
set was sampled with the same number of genes as the NTC or NTC subset (NR or NR coregulator) gene
set. The P-value of the significance of the overlap was estimated as the number of permuted gene sets that
contained at least as many genes present in the sex-biased gene set as in the NTC gene set, divided by the
total number of permutations.

RESULTS
Common and rare psychiatry-associated genetic variation is enriched with genes implicated in
nuclear receptor-mediated signaling
Whereas genetic variation in NR and NR coregulators, individually, has been associated with PDs in
association and linkage studies, the genetic risk profile of the NR transcriptome complex (NTC) as a whole
has not been systematically assessed at a large-scale, whole-genome level. More than 300 curated NR