Contains the codes for analyzing dN/dS values of neuron-specific genes and glial cell-specific genes across mammalian species.
Downloaded from http://hgdownload.soe.ucsc.edu/admin/exe/macOSX.x86_64/ on Nov 14, 2021
- run ./bash_scripts/mouse_protein_coding_dNdS.sh to download dN values, dS values, etc of protein-coding genes from Ensembl.
- Generate ./results/Ensembl98_mouse/mouse_protein_coding_genes.*.txt, One txt file for each of the 92 species with dN and dS values against mouse reference genome.
- run ./jupyter_notebooks/mouse.all_genes.ipynb to merge all 92 species together.
- Generate ./results/Ensembl98_mouse/mouse.92_species_dNdS.all_genes.tsv
- and ./results/Ensembl98_mouse/mouse.dNdS_stats.all_genes.tsv
- Visualize distribution of all genes' average dN/dS values, saved to ./figures/mouse.all_genes.*
- run ./jupyter_notebooks/mouse.celltype-expression.ipynb to calculate oligodendrocyte expression level from three subtypes and glial cell expression level from oligodendrocyte, microglia, and astrocyte
- Generate ./results/mouse.celltype-expression.tsv
- run ./jupyter_notbooks/mouse.celltype-specific-genes.ipynb to define cell type-specific genes.
- Generate ./results/mouse.celltype-specific_genes/all_barres/*.list.txt, one gene list for each cell type.
- Visualize the overlap between the lists with Venn diagrams, saved to ./figures/Venn_Diagrams/all_barres/*
- run ./jupyter_notebooks/mouse.celltype-specific_protein-coding_genes_w_dNdS.ipynb to merge protein-coding genes and dN/dS to celltype-specific genes
- Generate ./results/mouse.celltype-specific_genes/protein-coding_w_dNdS//.tsv, one per species per cell type, including averaged 92 species dN/dS
- run ./jupyter_notebooks/mouse.stats_and_figures.ipynb, calculate Mann Whitney stats for each pair of cell type for a couple of representative species as well as averaged dN/dS. Kruskal Wallis test was used for testing more than two cell types. Both annonical 0.05 p value threshold and a conservative 0.0001 p value threshold were used for visualizations. Figures with 0.0001 p value threshold were labeled with "conservative" in file name.
- Generate ./results/mouse.celltype-specific_genes/protein-coding_w_dNdS/MannWhitney/*.tsv
- and ./results/celltype-specific_descriptive_stats.xlsx
- and ./results/celltype-specific_inferential_stats.xlsx
- Visualize ./figures/MannWhitney/mouse_reference_genome/*.pdf
- and ./figures/MannWhitney/mouse.boxplots.pdf
- run ./jupyter_notebooks/mouse.GO-celltype-contingency.ipynb, analyze whether a few GOs are responsible for the neuron-specific genes with the lowest dN/dS, or if the pattern is universal throughout GOs.
- Generate ./results/celltype-specific_GOs.xlsx
- and ./results/NEAMO-specific_GOs.xlsx, which just takes out glia from the above file
- and ./results/GO_contingency_stats.xlsx
- Visualize ./figures/contingency_analysis/*.pdf
- run ./bash_scripts/mouse_organs.sh, merging organ-specific genes into a table
- Generate ./results/MGI_organs/MGI_ID-tissue_type.tsv and ./results/MGI_organs/MGI_ID_list.txt
- with ./results/MGI_organs/MGI_ID_list.txt, find Ensembl IDs of these genes
- Generate ./results/MGI_organs/Ensembl_ID-MGI_ID.tsv
- run ./jupyter_notebooks/mouse.organs.ipynb to generate figures comparing neuron-specific genes dNdS with organ-specific genes
- Generate ./figures/mouse.neuron_vs_organs.pdf and mouse.neuron_vs_organs_conservative.pdf
- and ./figures/mouse.organs.pdf
- run ./jupyter_notebooks/mouse.celltype-expressed-genes.ipynb, get list of genes expressed in each celltype with RPKM threshold of 1
- Generate ./results/mouse.celltype-expressed_genes/all_barres/*.list.txt
- Visualize ./figures/Venn_Diagrams/all_barres/*.png
- run ./jupyter_notebooks/mouse.celltype-expressed_protein-coding_genes_w_dNdS.ipynb
- Generate ./results/mouse.celltype-expressed_genes/protein-coding_w_dNdS//.tsv
- run ./jupyter_notebooks/mouse.celltype-expressed.stats_and_figures.ipynb
- Generate ./results/celltype-expressed_descriptive_stats.xlsx
- and ./results/celltype-expressed_inferential_stats.xlsx
- Visualize ./figures/MannWhitney/mouse_reference_genome/*.celltype-expressed.pdf
- and ./figures/mouse.celltype-expressed.boxplots.pdf
- run ./jupyter_notebooks/benchmarks.ipynb not cleaned up for github yet; check Herculano_Lab/Benchmark_Genes. For now use ./data/benchmarks.xlsx for analysis. Move that file to ./result once process is cleaned up and merged into Git.
- run ./bash_scripts/all_human_protein_coding_dNdS.sh to download human genes' orthologs information from 92 mammalian species (including mouse). This include pattern of ortholog and dN and dS values.
- Generate ./results/Ensembl98_human/human_protein_coding_genes.*.txt
- run ./jupyter_notebooks/human.all_genes.ipynb to merge all 92 species together.
- Generate ./results/Ensembl98_human/human.92_species_dNdS.all_genes.tsv
- and ./results/Ensembl98_human/human.dNdS_stats.all_genes.tsv
- Visualize distribution of all genes' average dN/dS values, saved to ./figures/human.all_genes.*
- run ./bash_scripts/chicken_protein_coding_dNdS.sh to download dN values, dS values, etc of protein-coding genes from Ensembl.
- Generate ./results/Ensembl98_chicken/chicken_protein_coding_genes.*.txt, One txt file for each of the 92 species with dN and dS values against mouse reference genome.
- run ./jupyter_notebooks/evo_distance.ipynb to look at the effect of evolutionary distance *
-
run ./jupyter_notebooks/mouse.gene_position.ipynb to get coordinates of 2,000 bp upstream 5' of each gene (for genes with ORF either + or - strand)
- Generate ./results/mouse.celltype-specific_genes/protein-coding_w_dNdS/promoter.*.tsv, one for each brain celltype
- and ./results/Ensembl100_mouse/promoter_position.tsv for all protein coding genes on mouse autosomes, sex chromosomes, and mitocondrial genes
-
run ../bash_scripts/parse_mouse_promoter_coord.sh to further format list of promoter coordinates for UCSC Genome Browser, which does not take header nor more than 1000 lines at a time
- Generate ./results/mouse.celltype-specific_genes/protein-coding_w_dNdS/promoter.*aa or ab if that celltype has more than a thousand promoters.
-
Use the coordinate to reterive phastCons stats from UCSC genome browser's table browser with GRCm38 mouse reference genome, with track phastCons60way (July 28th, 2020). Compile differenct cell types into an excel file by hand.
- Generate ./results/mouse.celltype-specific_genes/protein-coding_w_dNdS/phastCons/phastcons-celltype.xlsx
- run ./jupyter_notebooks/Bayes_neuron_gene_venn.ipynb
- get venn diagram between their neuronal genes (hPSD or non-hPSD) and my mouse neuron-specific genes
- mostly non-overlapping
- run ./jupyter_notebooks/mouse_adult_brain.Saunders_2018_annotation.ipynb
- run ./jupyter_notebooks/mouse_adult_brain.Saunders_2018_Cell_Type.ipynb
- run ./jupyter_notebooks/exc_inh.mouse_adult_brain.Saunders_2018.ipynb
- test whether excitatory and inhibitory neuron-specific genes are different in dN/dS values
- run ./jupyter_notebooks/neuron_subtypes.mouse_adult_brain.Saunders.ipynb
- test whether the 500+ subclusters, aka detailed neuronal, glial, and vasculature cell types, are different in dN/dS
- this is to test whether my original result that neuron-specific genes are low in dN/dS is due to these genes ubiquitously expressed across neuronal subtypes.
- run ./jupyter_notebooks/mouse.zeisel_2015.ipynb
- Confirming that neuron-specific genes have lowest dN/dS with Linnarson lab's 2015 single cell RNA-seq data
- statistical results for Mann Whitney U test recorded in ../results/single-cell/zeisel_2015/mwu-stats.tsv
- descriptive stats in ../results/single-cell/zeisel_2015/nine_type_median_dNdS_CI95.xlsx
- Also generated figures visualizing the differences in dN/dS
- run ./jupyter_notebooks/mouse.TaxonomySymbol.zeisel_2018.ipynb
- Getting list of TaxonomySymbol from aggregate level 5 loom file's column attributes
- output ./data/single-cell/zeisel_2018/TaxonomySymbol_list.txt
- this file is then mannually combined with values from Taxonomy_group column in ./data/single-cell/zeisel_2018/mmc3.xlsx
- which is stored in ./data/single-cell/zeisel_2018/Taxonomy_Symbol-to-Group_Name.xlsx
- Labeling ./data/single-cell.zeisel_2018/mmc3.xlsx mannually
- For CNS neuron clusters (Taxonomy_group is in [Telencephalon projecting excitatory neurons, Dentate gyrus granule neruons, Telecephalon projecting inhibitory neurons, Olfactory inhibitory neurons, Telencephalon inhibitory interneurons, Cholinergic and monoaminergic neurons, Peptidergic neurons, Spinal cord inhibitory neurons, Di- and mesencephalon excitatory neurons, Di and mesencephalon inhibitory neurons, Hindbrain neurons, Cerebellum neurons]):
- Clusters with Taxonomy group in [Cholinergic and monoaminergic neurons, peptidergic neruons]:
- label as mix_neuron
- Clusters described as neuroblasts in description:
- label as neuroblasts
- GABAergic Clusters with Trinarization detecting VGLUT and/or Acetylcholine, or Glutamatergic Clusters with Trinarization detecting GABA in the comment column:
- label as mix_neuron
- Clusters with Neurotransmitter labeled as just 1) GABA, 2) GABA and Glycine, 3) GABA and Nitric Oxide, or 4) GABA, Nitric Oxide, and Glycine, and without any other neurotransmitter:
- label as inh_neuron
- Clusters with Neurotransmitter labeled as a combination of VGLUT, with or without Nitric Oxide, and without any other neurotransmitter:
- label as exc_neuron
- Clusters with Taxonomy group in [Cholinergic and monoaminergic neurons, peptidergic neruons]:
- Taxonomy_group == Enteric neurons
- label as enteric_neuron
- Taxonomy_group is in [Sympathetic noradrenergic neruons, Sympathetic cholinergic neurons, Peripheral sensory peptidergic neurons, Peripheral sensory peptidergic neruons, Peripheral sensotry neurofilament neruons, Peripheral sensory non-peptidergic neurons]:
- label as pns_neuron
- Taxonomy_group is in [Oligodendrocytes, Oligodendrocyte precursor cells, Ependymal cells, Astrocytes, microglia]:
- label as glia
- Taxonomy_group is in [Choroid epithelial cells, Subcommissural organ hypendymal cells, Dentate gyrus radial glia-like cells, Subventricular zone radial glia-like cells, Olfactory ensheathing cells, Vascular and leptomeningeal cells]:
- label as other
- Taxonomy_group is in [Schwann cells, Satellite glia]:
- label as PNS_glia
- Taxonomy_group == Enteric glia:
- label as enteric_glia
- Taxonomy_group is in [Vascular smooth muscle cells, Pericytes]:
- if cluster is VECA (description says endothelial cells but Taxonomy is smooth muscle and clustered together with PER3 pericytes):
- label as other
- else:
- label as excitable
- if cluster is VECA (description says endothelial cells but Taxonomy is smooth muscle and clustered together with PER3 pericytes):
- Taxonomy_group == Vascular endothelial cells
- label as endothelia
- Taxonomy_group == Perivascular macrophages
- label as immune
- output as linhe_class column of expanded.mmc3.xlsx
- For CNS neuron clusters (Taxonomy_group is in [Telencephalon projecting excitatory neurons, Dentate gyrus granule neruons, Telecephalon projecting inhibitory neurons, Olfactory inhibitory neurons, Telencephalon inhibitory interneurons, Cholinergic and monoaminergic neurons, Peptidergic neurons, Spinal cord inhibitory neurons, Di- and mesencephalon excitatory neurons, Di and mesencephalon inhibitory neurons, Hindbrain neurons, Cerebellum neurons]):