Isosceles paper

This repository contains scripts and data analysis for the Isosceles paper, enabling the replication of data, analyses, and plots featured in the paper. The repository is divided into two main sections:

Section #1. Analysis of Quantifications

In this section, you can access singularity images and benchmark program outputs required to run our scripts in the reports folder, reproducing the figures and plots presented in the paper. Each directory includes a script to download the results from running all programs on the corresponding benchmark. You can execute these scripts using the provided commands below. For more details on the exact commands used to generate the data, please refer to the Supplementary Section at the bottom or check the brief summary in the Benchmark_commands.md document.

Download and prepare the Singularity images:

cd singularity
bash run.sh
cd ..

Prepare reference genome and annotations data:

cd reference_data
bash run.sh
cd ..

Download the data needed to create the reports:

cd simulated_bulk_analysis
bash download_report_data.sh
cd ..
cd simulated_sc_analysis
bash download_report_data.sh
cd ..
cd simulated_ovarian_analysis
bash download_report_data.sh
cd ..
cd nanopore_bulk_analysis
bash download_report_data.sh
cd ..
cd nanopore_sc_analysis
bash download_report_data.sh
cd ..
cd sirv_analysis
bash download_report_data.sh
cd ..
cd sequin_analysis
bash download_report_data.sh
cd ..
cd pacbio_analysis
bash download_report_data.sh
cd ..

Then you can run the desired reports to generate figures - this can be done manually in RStudio and Jupyter, or using the following commands (the latter requires Singularity to be installed):

cd reports
bash run.sh
cd ..

Section #2. Mouse E18 Brain scRNA-Seq (Lebrigand et al.) Data Analysis

The analysis of the Lebrigand et al. data from the paper can be replicated using the code in the mouse_E18_analysis directory. You can download the input data for the analysis (including scRNA-Seq BAM files prepared using the Sicelore workflow described in Part #1), or start from any downstream step, as RDS files containing data necessary for further analysis are provided for every script.

Download and prepare the Singularity images:

cd singularity
bash run.sh
cd ..

Prepare input data for the analysis:

cd mouse_E18_analysis
bash download_data.sh
cd ..

The mouse_E18_analysis directory contains the following scripts for generating the figures and tables:

00_run_isosceles.R
01_scrnaseq_analysis.R
02_run_dexseq_intra.R
03_run_dexseq_inter.R
04_psi_heatmap_intra.R
05_psi_heatmap_inter.R
06_other_plots.R
07_isoswitch_analysis.R

You can run these scripts manually in RStudio, or using the following commands (the latter requires Singularity to be installed):

cd mouse_E18_analysis
bash run.sh
cd ..

Supplementary Section. (Optional) Re-generating Quantifications from Raw Reads

We also provide the scripts to re-run all the programs to re-generate quantifications used in Part #1 reports (provided you download the input data). This is more involved, and requires significant computation and memory resources. While Isosceles is fairly efficient and can be run on most systems, not all software benchmarked here is, and so most steps were run on a cluster with 20 CPUs and 200 GB of RAM. While we provide environments for each software suite, we note that although this runs cleanly on our system, some programs may take some troubleshooting on other systems to get working.

First prepare the software and environments using:

cd software
bash run.sh

Prepare the Singularity images:

cd singularity
bash run.sh
cd ..

Prepare reference genome and annotations data using:

cd reference_data
bash run.sh
cd ..

Prepare input data for the analysis using:

cd input_data
bash run.sh
cd ..

Then in each of the following directories you can follow the run.sh steps:

cd simulated_bulk_analysis
bash run.sh
cd ..
cd simulated_sc_analysis
bash run.sh
cd ..
cd simulated_ovarian_analysis
bash run.sh
cd ..
cd illumina_sc_analysis
bash run.sh
cd ..
cd nanopore_bulk_analysis
bash run.sh
cd ..
cd nanopore_sc_analysis
bash run.sh
cd ..
cd sirv_analysis
bash run.sh
cd ..
cd sequin_analysis
bash run.sh
cd ..
cd pacbio_analysis
bash run.sh
cd ..

Finally to prepare the data for the analysis scripts in reports, as used in Part #1, you can run the prepare_report_data.sh (or prepare_report_data.R, depending on the directory) scripts:

cd simulated_bulk_analysis
singularity exec ../singularity/isosceles.sif Rscript prepare_report_data.R
cd ..
cd simulated_sc_analysis
bash prepare_report_data.sh
cd ..
cd simulated_ovarian_analysis
bash prepare_report_data.sh
cd ..
cd nanopore_bulk_analysis
bash prepare_report_data.sh
cd ..
cd nanopore_sc_analysis
bash prepare_report_data.sh
cd ..
cd sirv_analysis
bash prepare_report_data.sh
cd ..
cd sequin_analysis
bash prepare_report_data.sh
cd ..
cd pacbio_analysis
bash prepare_report_data.sh
cd ..