Long-read RNA sequencing identifies region- and sex-specific C57BL/6J mouse brain mRNA isoform expression and usage
2024-01-12
bioRxiv Preprint:
Emma F. Jones, Timothy C. Howton, Victoria L. Flanary, Amanda D. Clark, and Brittany N. Lasseigne.
The University of Alabama at Birmingham, Heersink School of Medicine, Department of Cell, Developmental and Integrative Biology
To assess differences in AS across the cerebellum, cortex, hippocampus, and striatum by sex, we generated and analyzed Oxford Nanopore Technologies (ONT) long-read RNA sequencing (lrRNA-Seq) C57BL/6J mouse brain cDNA libraries. From >85 million reads that passed quality control metrics, we calculated differential gene expression (DGE), differential transcript expression (DTE), and differential transcript usage (DTU) across brain regions and by sex. We found significant DGE, DTE, and DTU across brain regions, and that the cerebellum had the most differences compared to the other three regions. Additionally, we found region-specific differential splicing between sexes, with the most sex differences in DTU in the cortex and no DTU in the hippocampus. We also report on two distinct patterns of sex DTU we observed, sex-divergent and sex-specific, that could potentially help explain sex differences in the prevalence and prognosis of various neurological and psychiatric disorders in future studies. Finally, we built a Shiny web application for researchers to further explore the data.
You can also visit our interactive Shiny Application that accompanies this analysis:
https://lasseignelab.shinyapps.io/mouse_brain_iso_div/
The sequencing data generated in this project is available on GEO at GSE246705
## src/preprocessing
## ├── 00_run_nanoseq.sh
## ├── 02_get_genome_annotations.sh
## ├── 03_create_isoform_fa.sh
## ├── 16_dataset_overview_figure1.Rmd
## └── README
Script 00: preprocessing/00_run_nanoseq.sh - This script can be skipped if you are working from counts files. Does not require a docker. You need raw input fasta files, a genome file, and genome annotation file. I used GENCODE mouse release M31.
Script 02: preprocessing/02_get_genome_annotations.sh - You need genome annotations available for gffread to run, but you also need genome annotations to run the nanoseq pipeline so frankly you should already have them somewhere and could just move them. You do need to get genome annotations if you didn’t run the nanoseq pipeline and are just using counts, so I am keeping it as script 2. It took me 20 minutes to download.
Script 03: preprocessing/03_create_isoform_fa.sh - This script depends on script 02. It is a single command but needs to be run in a docker (either RStudio docker I made has gffread) or on a local machine with gffread.
Script 16: preprocessing/16_dataset_overview_figure.Rmd - The purpose of this script is to provide a dataset overview to serve as manuscript figure 1. It is dependent on scripts 00 and 01 to get sample metadata. Must run in github docker 1.7 or above.
## src/dtu_analyses
## ├── 01_calculate_cpm.R
## ├── 04_pca_eda.Rmd
## ├── 05_dtu_region_region.Rmd
## ├── 06_dtu_region_others.Rmd
## ├── 07_dtu_region_sex.Rmd
## ├── 08_dtu_neuro_diseases.Rmd
## ├── 09_dtu_isoform_switching.Rmd
## ├── 14_protein_domain_info.Rmd
## ├── README
## ├── functions.R
## └── size_power.R
dtu_analysis/functions.R - This is a function script for all Rmd files in the dtu directory.
Script 01: dtu_analyses/01_calculate_cpm.R - This script is very short and depends on having transcript counts available either by downloading or running the nanoseq pipeline. Run this script in the github.1.2 docker or above. Takes less than one minute.
Script 04: dtu_analyses/04_pca_eda.Rmd - This script is for exploratory data analysis and PCA. It depends on script 01. Run this script in the github.1.2 docker or above. If data looks bad, do not proceed to script 05, but it is technically independent. This script takes 3 minutes to run.
Script 05: dtu_analyses/05_dtu_region_region.Rmd - This script depends on the outputs from script 01 which are read as an RDS file. This script also depends on script 03 - the gffread script and the functions script. Run in github.1.2 docker so you have the most up-to-date version of the package. This script takes about 10 minutes to run.
Script 06: dtu_analyses/06_dtu_region_others.Rmd - This script is meant for DTU analysis comparing one brain region to all others. This script depends on scripts 01-03 and the functions script. Run this in the github.1.2 docker or above. Takes about 10 minutes to run.
Script 07: dtu_analyses/07_dtu_region_sex.Rmd - This script is meant for DTU analysis across sexes within brain regions. This script depends on scripts 01-03 and functions script. Run this in the github.1.2 docker or above. Takes about 20 minutes to run (I think).
Script 08: dtu_analyses/08_dtu_neuro_diseases.Rmd - This script is for comparing brain-region-specific DTU genes to known Alzheimer’s disease, psychiatric disorder, and CPAM case genes. Run this in the github.1.2 docker or above. It is dependent on scripts 01 - 06 and the functions script. It takes 4 minutes to run.
Script 09: dtu_analyses/09_dtu_isoform_switching.Rmd - This script is for adding open reading frames to the existing switchlist objects. It is also for plotting and saving switch plots, which show the significant isoform switching events. Run this in the github.1.2 docker or above. It is dependent on scripts 01-07 and the functions script. This script also takes 36 minutes to run.
Script 14: dtu_analyses/14_protein_domain_info.Rmd - This script depends on scripts 00-13, and also includes a shell script for running perl code. The purpose of this script is to extract nucleotide and amino acid sequences and run pfam to annotate the protein domains. This script needs to run in the pfam-tagged docker image (pfamscan.1.1) with perl!
## src/de_analyses
## ├── 10_DESeq2_region_region.Rmd
## ├── 11_DESeq2_region_others.Rmd
## ├── 12_DESeq2_region_sex.Rmd
## ├── 13_incorporate_de_results.Rmd
## ├── 15_compare_DTU_DGE.Rmd
## ├── 17_compare_region_pairwise_results.Rmd
## ├── 18_enrichment_analysis.Rmd
## ├── 19_create_fig_2.Rmd
## ├── 20_compare_enrichment_results.Rmd
## ├── README
## ├── de_functions.R
## └── stacked_barplot_functions.R
de_analyses/de_functions.R - This is a function script for all Rmd files in the de directory.
de_analyses/stacked_barplot_functions.R - This is a function script for specifically building stacked barplots for figure 2.
Script 10: de_analyses/10_DESeq2_region_region.Rmd - This script is dependent on script 00/01 or having gene and transcript level count data with metadata in your /data/ directory and the de functions script. The purpose of this script is to run DESeq2 across brain regions at the gene and transcript level. It takes less than 5 minutes to run. Must run in github docker 1.7.
Script 11: de_analyses/11_DESeq2_region_others.Rmd - This script is dependent on script 00/01 or having gene and transcript level count data with metadata in your /data/ directory and the de functions script. The purpose of this script is to run DESeq2 for each brain region compared to an aggregate of other brain regions at the gene and transcript level. It takes about 10 minutes to run. Must run in github docker 1.7.
Script 12: de_analyses/12_DESeq2_region_sex.Rmd - This script is dependent on script 00/01 or having gene and transcript level count data with metadata in your /data/ directory and the de functions script. The purpose of this script is to compare expression at the gene and transcript level across sexes within brain regions. It takes less than 10 minutes to run. Must run in github docker 1.7.
Script 13: de_analyses/13_incorporate_de_results.Rmd - This script is fully dependent on scripts 01-09 and the de_analysis DESeq2 scripts 10-12 and the de functions script. The purpose of this script is to wrap in all DESeq2 significance values into my isoformSwitchAnalyzeR objects and subsequently, plots. This enables us to plot them all with significance values for DGE, DTE, and DTU. This script takes 12 minutes to run. Must run in github docker 1.7.
Script 15: de_analyses/15_compare_DTU_DGE.Rmd - This script depends on scripts 00-14. The purpose of this script is to compare genes with differential gene expression, differential transcript usage, and differential transcript expression. Must run in docker 1.7 to have all packages.
Script 17: de_analyses/17_compare_region_pairwise_results.Rmd - The purpose of this script is to determine how many DTU genes or DEGs are specific to a study design. I used Jaccard Similarity to see how similar the results are to each other and create supp figure 3. Must run in github docker 1.7.
Script 18: de_analyses/18_enrichment_analysis.Rmd - The purpose of this script is to perform functional enrichment analysis with the package gprofiler2 on the DGE and DTE genes. This script is fully dependent on scripts 00-12. Must run in github docker 1.7.
Script 19: de_analyses/19_create_fig_2.Rmd - The purpose of this script is to create supplementary figure 2. This script is fully dependent on scripts 00-12. Must run in github docker 1.7.
Script 20: de_analyses/20_compare_enrichment_results.Rmd - The purpose of this script is to get real statistics and values for directly comparing functional enrichment analysis results. It is dependent on scripts 00-18. Run in github docker 1.7.
## src/shiny_support
## ├── README
## └── calculate_cpm_convert_ids.R
shiny_support/calculate_cpm_convert_ids.R - This script enables the gene IDs for the Shiny App to be either ENSEMBL IDs or gene symbols and does some other needed reformatting of the data for the application to work.
## src/shiny_app
## ├── README
## ├── app.R
## ├── complete_switchlists
## │ ├── region_all_list_orf_de_pfam.Rds
## │ ├── region_region_orf_de_pfam.Rds
## │ └── region_sex_list_orf_de_pfam.Rds
## ├── data
## │ ├── combined_cpm.Rds
## │ └── sample_collection_metadata.Rds
## ├── rsconnect
## │ └── shinyapps.io
## │ └── lasseignelab
## │ └── mouse_brain_iso_div.dcf
## ├── server.R
## ├── ui.R
## └── www
## ├── DTU_example.png
## ├── code.js
## ├── favicon.ico
## ├── graphical_abstract.png
## ├── logo_only.png
## └── style.css
shiny_app/app.R - This script runs the entire Shiny application.
shiny_app/server.R - The purpose of this script is to run the back-end code needed for generating plots and data for the Shiny application.
shiny_app/ui.R - The purpose of this script is to run the user interface for shiny web application.
All intermediate data generated by the following analyses can be found at the following links:
Data:
Docker Images:
GitHub Repository:
We performed all analyses in docker with R version 4.3.1. The repository with all the Docker images used for this project can be found on Docker Hub at emmafjones/mouse_brain_iso_div. While individual docker image tags are noted in every script, we provide an overview of analyses and their associated tags below:
| Tag | Associated Analyses |
|---|---|
| github.1.7 | All scripts 01-13, 15-20 |
| pfamscan.1.1 | Script 14 |
What is Happening in the Lasseigne Lab?
This work was supported in part by the UAB Lasseigne Lab funds, R00HG009678, and the UAB Pittman Scholar Award.
We acknowledge all current and past members of the Lasseigne Lab for their thoughtful feedback, especially Tabea M. Soelter, Jordan H. Whitlock, Vishal H. Oza, and Elizabeth J. Wilk. We would like to thank the UAB Biological Data Sciences (UAB-BDS) core for discussions during office hours and institutional support of the Cheaha configuration for nf-core pipelines and docker/singularity container documentation. All figures and cartoons were assembled with BioRender.
This repository is licensed under the MIT License, see LICENSE documentation within this repository for more details.
