Code and data for "Inferred divergent gene regulation in archaic hominins reveals potential phenotypic differences"
https://doi.org/10.1038/s41559-019-0996-x
We anticipate that the most common quest that would bring people here is searching for the list of DR genes dicussed
in the majority of the paper. To get the full list of DR genes in the Altai Neanderthal discussed in figures 2-4, look in
./data/altai_original_pvalues_2sided.txt. This is a gene x tissue matrix containing empirical p-values calculated vs.
1000 Genomes. A gene is DR in a particular tissue if p=0, and "NA" indicates that the gene was not succesfully modeled in
that tissue. If you want specifically the DR GWARRs you look at altai_original_nonIntro_DR_genes.txt. If you want to
know what tissues they were DR in, you can join those ids with the p-value file.
If you want DR genes to compare between archaic hominins, look in ./data/altai_update_pvalues_2sided.txt,
./data/denisovan_update_pvalues_2sided.txt, and ./data/vindija_pvalues_2sided.txt and again identify genes for
which p=0 in a given tissue.
Further description of files in this repo:
comp_pop.jl
calculates empirical p-values, runs PCA and clustering analyses
requires population predictions and archaic predictions or pre-computed p-values
comp_inds.jl
comparisons of predicted expression between archaic individuals
requires archaic predictions
gene_overlap.jl
intersects gene models and SNPs or regions of interest
requires lists of regions of interest
human_dr.jl
for a set of human IDs, calculates empirical P-value and call DR genes
intro_individuals.jl
for each gene, identifies individuals with certain alleles within 1Mb
metric_summary_by_gene.jl
generalized version of recombination_rate_map.jl; compares a metric (i.e. gene density) across sets of genes
model_weights.jl
stats for comparing model weights between sets of SNPs
requires model databases, SNP sets
neanderthal_predixcan_plots.ipynb
Jupyter notebook for plotting
PrediXcan.py
code for performing PrediXcan predictions on genetic information.
adapted from script downloaded here on May 19, 2017
recombination_rate_map.jl
calculates recombination rate and summarizes for gene regions, then compares R.R. across gene sets.
*predExp.txt
archaic prediction files-- genes x tissue matrix containing PrediXcan output for each gene model. altai_original
indicates original genome called for Altai neanderthal, used for most analyses. altai_update is the updated one used
for comparisons to other hominins.
*pvalues_2sided.txt
archaic 1kG empirical p-values-- genes x tissue matrix. matched to predExp file. A gene is DR in a tissue
if this p-value equals zero. NA indicates that the gene was not successfully modelled in that tissue.
altai_disgenet.txt
Disgenet enrichments for DR GWARRS. corresponds to fig. 2d.
altai_dr_phewas.txt
phenotype associations with differential regulation of DR GWARRs. Corresponds to figure 3 and Table 1.
altai_desert_genes.txt
genes with at least 1 SNP in any of their tissue models that overlaps an introgression desert. Ensembl ids.
altai_intro_genes.txt
genes with at least 1 SNP in any of their tissue models that is on an introgressed haplotype (ie are non-GWARRs).
Anti-join this list and the genes in predixcan_model_summary.txt to get all GWARRs.
altai_intro_genes_names.txt
human-readable names of the genes in the file above.
altai_original_nonIntro_DR_genes.txt
ensembl ids for all DR GWARRs. locations plotted in figure 2b.
altai_original_missingModels.txt
all genes for which the Altai genome was missing all SNPs in at least one tissue model. These were excluded
from analyses.
gene-dersnp_intersection.txt
genes by ensembl ID and counts of archaic-derived SNPs within 1Mb of them
gene_region_neanSNPs.txt.gz
for each gene, list of introgressed SNPs (by rsID) within 1 Mb (ie the region considered by PrediXcan).
DR_crossSpeces_GO.tgz
cross-species GO analyses- overlap of DR genes. Corresponds to figure 5b.
diff_crossSpeces_GO.tgz
cross-species GO analyses- absolute difference in regulation. corresponds to figure 5c.
gencode.v12.annotation.bed.gz
hg19 gene locations. A gene's location is considered to be all bases covered by at least 1 version of the gene.
predixcan_model_summary.txt
summary of all predixcan models in all tissues. models described by ensembl id, and includes model performance
(R2 for observed/predicted in training), number of snps in the model, number of those SNPs that were
missing in Altai_original, number of introgressed SNPs in the model (if 0 it's a GWARR), number of SNPs shared
between AMH and Altai, and the empirical p-value for divergent regulation (same as in
altai_original_pvalues_2sided.txt)