Welcome to Red-C RNA-DNA interactome mapping with nextflow.
Red-C is a novel method for genome-wide detection of RNA-DNA interactions in chromatin. The details are in our NAR paper from 09 July, 2020:
Gavrilov, A.A., Zharikova, A.A., Galitsyna, A.A., Luzhin, A.V., Rubanova, N.M., Golov, A.K.,
Petrova, N.V., Logacheva, M.D., Kantidze, O.L., Ulianov, S.V., Magnitov, M.D., Mironov, A.A., and Razin S.V. 2020.
Studying RNA–DNA interactome by Red-C identifies noncoding RNAs associated with various chromatin
types and reveals transcription dynamics.
Nucleic Acids Research.
The initial version of the pipeline (RedClib) was implemented in Python. It was hard to scale. From now on, we support the nextflow pipeline redc-nf.
Pros:
- You can easily change the parameters of run.
- The data processing is parallel and scalable.
- No need to re-run from scratch due to nextflow resume functionality.
- You can switch off unneded processing steps.
Cons: none detected.
- You need a system supporting nextflow pipelines. For example, you can create conda environment with nextflow, which will contain redc-nf dependencies:
conda env create -n redc-nf -f environment.yml
conda activate redc-nf
conda install -c bioconda nextflow
- Run test example:
nextflow redc.nf -params-file project.yml
This example should take up to 1 minute.
you may want to check that the results are the same as we provide in test/
folder.
- Adjust
project.yml
for your needs. You may vary libraries, input files, genome, RNA annotation, restriction enzymes, filters, output files and their content, chunk size for splitting the data, oligos and trimming parameters.
project.yml
has inline documentation that should explain the details.
- Check out the results of mapping and annotation of your RNA-DNA contacts. In case of troubles, visit and improve our issues page.
Also, take a look at the outdated description of RedClib, but might provide some hints on the inner structure of redc-nf pipeline:
- 30-01-2021: Nota that hisat2 output with defalt annotation of restriction sites depends on the reads order.
See issue #222 of hisat on GitHub.
Thus the results of redc-nf may vary for different chunk sizes.
If you want to have the same results for different chunk sizes, add
--no-temp-splicesite
option and remove--novel-splicesite
annotation. This reduces the performance of splice sites detection and makes the results reproducible for runs with different chunk sizes.