Recovering significant 10xGenomics single-cell RNA-seq signal through improved annotation with Oxford Nanopore (ONT) bulk long reads. Applied on chick embryo neuro-epithelial progenitors at 66h of development.

GitHub repository started on 2021-04-02.

• data - Data used for the analyses (e.g. BAM, RDS, tsv files, ...)
• docs - Rendered analysis reports (as html files) and figures generated by the notebook
• env - Code used to create the conda environment for this study (e.g. bash script) and the corresponding YAML file.
• figs - Figures created before the notebook creation and used as input
• rmd - R Markdown analysis files
• src - Reusable code (e.g. functions)

Files found at the root of the repository are of general purpose:

• _bookdown.yml - Configuration YAML file for the notebook
• _build.log - Log file (stdout and stderr) generated when running the analysis
• _build.sh - Bash script to run the analysis
• _deploy.sh - Bash script to deploy the analysis on GitHub
• index.Rmd - R Markdown file used for the setup of the analysis (load libraries, define variables, paths...)
• README.md - This file
• _workflowr.yml - Workflowr configuration file

Current input data files include:

- notebook		# javascript code for bookdown output (do not touch)
- processed		# output generated when running the notebook
- raw			# raw data (input)
  - references
    - annotations
	  - ensembl
	  - ncbi
	  - ucsc
  - rna-seq
    - single-cell
    - long-read

Current notebook files include:

• 01-Impact-ref-annotation-scRNA.Rmd

Here we explore the discrepancies between the references annotations (Ensembl and RefSeq) and their impact on common scRNA-seq analyses.

• 02-Incomplete-annotations-induce-signal-loss.Rmd

We study here the loss in scRNA-seq signal (e.g. genes) due to significant deficiencies in the reference annotation, specially in the 3' UTRs annotations.

• 03-Approaches-to-improve-transcriptome-with-Long-Reads.Rmd

We compare various tools dedicated to transcriptome reconstruction in bulk RNA-seq (StringTie2, scallop), a dedicated signal detection approach and broad 3' UTR extension and apply them to scRNA-seq data and ONT bulk long reads.

• 04-Impact-reannotation-scRNA.Rmd

We assess the impact of our various reannotations on common scRNA-seq analyses.

• 05-Validation-of-novel-genes-with-scRNA.Rmd

We evaluate the ability of our approach to identify novel genes and use scRNA-seq analyses as a filter to highlight genes of biological interest in chick embryo neuro-epithelial progenitors.

• 06-A-tool-and-pipeline-to-improve-annotation-for-scRNA.Rmd

Description and recommendations to use our pipeline on other scRNA-seq data.

• 07-Session-info.Rmd

Session info output.

Current code files include:

- analysis
- pipeline
- preprocessing
- utilities

To run the notebook and create the corresponding html files, you have two options:

• In RStudio, click the knit button (you may need to change the knit directory)
• In a linux terminal, run the script _build.sh with the command:

bash _build.sh

The output will be stored in the docs folder.