ONTflow

Workflow for analysing sequencing data from Oxford Nanopore Technologies based on Snakemake.

Functionalities

1. Transcripts and gene abundance estimation.
2. Gene fusion detection.

Overview of the workflow

ONTflow is a workflow for analyzing long-read sequencing data from Oxford Nanopore Technologies (ONT) sequencer based on Snakemake which enables the workflow reproducible and scalable. It is also based on Anaconda which simplifies the installation of all required tools and also ensures reproducibility.

The origin intention of developing ONTflow is to analyze in-house patients' sequencing data which is sensitive and all analyses need to be executed within TSD which is a closed system, and Anaconda cloud is therefore not accessible. In that case, tools have to be prepared one by one manually.

ONTflow takes FASTQ files generated from the sequencer as input, and generated mainly three outputs: abundances of transcripts and genes, and detected gene fusions.

With Graphviz installed, the workflow can be visualized by snakemake --rulegraph -s workflow.smk | dot -Tsvg > workflow.svg

Installation

For TSD users

Most tools are ready on TSD

Most tools are either already available on TSD and can be loaded directly, or already installed at a shared directory, as can be seen in the SLURM workscript file at ./scripts/workscript_nanopore_main_smk.slurm.

Install three R packages

But the R packages need to be installed by each user, since they can only be installed at personal folder as required by TSD.

Firstly, load the specific R with module load R/3.5.0, and then get into R by command R. Then install the packages inside there:

install.packages('hash')

install.packages('tximport')

But it's a bit tricky to install "stringr" which requires dependency "stringi". You have to install "stringi" with install.packages("/tsd/shared/R/misc/stringi_1.4.6.tar.gz", repos = NULL, type = "source") (more information here: https://www.uio.no/english/services/it/research/sensitive-data/help/software/r/using-R.html#toc6

Then you can install "stringr" normally:

install.packages('stringr')

Install one Python package

Another tool that needs to be installed manually (also by each user) is pycoQC which is the tool to do quality control of the sequencing.

Firstly, load the specific Python with module load python3.gnu/3.7.3, and then install pycoQC with:

pip3 install --user pycoQC

The software will automatically be installed to your own local directiry ~/.local/bin. That's why you see this line in workscript: export PATH=~/.local/bin:$PATH so that SLURM knows where to find pycoQC.

For Linux users

Install Miniconda

Download Miniconda installer from here: https://docs.conda.io/en/latest/miniconda.html

Install it to your laptop or server.

Install ONTflow

Download the repository from here: https://github.com/zhxiaokang/ONTflow/archive/refs/heads/main.zip and unzip it.

If you have Git installed on your machine, a better way to download the repository is to use git: $ git clone https://github.com/zhxiaokang/ONTflow.git

With that, you can easily keep yourself updated with the latest version using the command git pull.

Open your terminal in the directory of the repository ONTflow and run the following command to set up the environment: $ conda env create -n ontflow -f config/env.yaml

Then activate the environment: $ conda activate ontflow

But Guppy is not open source, and it is only available for ONT customers. For ONT customers, please download and install Guppy manually from the ONT community. And add the installed path to the environment: $ export PATH=/path/to/guppy:$PATH

For Mac and Windows users

A Docker image will be provided later.

Tips for Windows users: For some Windows operating systems, Windows Subsystem for Linux (WSL) is available. To install it, refer to https://docs.microsoft.com/enus/windows/wsl/install-win10 and then follow the instruction for Linux users above.

Setting up

The Nanopore sequencer generates a lot of FASTQ files, the most interesting ones are in folder fastq_pass. The FASTQ files are stored in folders named like barcode*. Creat a file named barcode_index.txt and list all the folders' name there (one name each row). For example:

barcode01
barcode02
barcode03
barcode04
barcode05
barcode06
barcode07
barcode08
barcode09
barcode10
barcode11
barcode12

In the case of Genome Biology group OUS, go to config/config_nanopore_ous_p19.yaml and modify these 4 parameters accordingly: BARCODE_FILE, SEQ_SUMMARY, DIR_FASTQ, and DIR_OUTPUT. And for other users, go to config/config.yaml and modify the parameters accordingly.

Run the workflow

For TSD users

Go to the directory script/ and run sh runscript_nanopore_main_smk.sh from there.

For other users

Go to the directory script/ and run snakemake --cores 12 -s nanopore_main.smk from there.

Known issues

snakemake may report some errors due to its own version conflict, eg. the version 3.13.3 does not need the label directory at output.