https://github.com/cmdoret/acastellanii_hybrid_assembly
Hybrid genome assembly of Acanthamoeba castellanii. Combining long reads, Hi-C and shotgun.
https://github.com/cmdoret/acastellanii_hybrid_assembly
Last synced: about 1 year ago
JSON representation
Hybrid genome assembly of Acanthamoeba castellanii. Combining long reads, Hi-C and shotgun.
- Host: GitHub
- URL: https://github.com/cmdoret/acastellanii_hybrid_assembly
- Owner: cmdoret
- License: gpl-3.0
- Created: 2019-07-18T18:18:58.000Z (almost 7 years ago)
- Default Branch: master
- Last Pushed: 2021-10-01T17:02:04.000Z (over 4 years ago)
- Last Synced: 2025-02-17T23:47:35.849Z (over 1 year ago)
- Language: Python
- Size: 175 KB
- Stars: 1
- Watchers: 2
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
### Acanthamoeba castellanii Hybrid assembly pipeline
[](https://doi.org/10.5281/zenodo.5507417)
This repository contains an automatic and pipeline for the genome assembly of Acanthamoeba castellanii. The different steps of the pipeline are illustrated below.
The analysis is implemented by combining the snakemake workflow system with [singularity containers](https://snakemake.readthedocs.io/en/stable/snakefiles/deployment.html#singularity) and conda environments. Each pipeline step is encapsulated and dependencies are managed automatically.
All input reads are downloaded automatically from SRA, based on their accession numbers defined in `units.tsv`. A frozen copy of this repository and its output files are available on the associated [Zenodo record](https://doi.org/10.5281/zenodo.5507417).
#### Dependencies
* python >=3.7
* singularity >= 3.2
* conda >= 4.8
Python packages:
* snakemake >= 5.5
#### Usage
The pipeline can be started using:
```bash
snakemake --use-conda -j4
```
or
```bash
snakemake --use-singularity -j4
```
to use singularity instead of conda environments.
#### Pipeline steps
The pipeline uses 3 types of input data:
* shotgun Illumina reads
* Hi-C Illumina reads
* Oxford Nanopore long reads
All those reads are automatically downloaded from SRA when running the pipeline.
The initial assembly is performed with long reads only using Flye. The short reads are then used to polish this assembly using HyPo. The Hi-C scaffolding is done using instagraal, followed by instagraal-polish to fix errors introduced by instagraal.
>TODO: Add quast report at the end of the pipeline
Each rule requiring a third party software pulls a standalone container hosted on dockerhub or quay to work in an isolated environment with a fixed version of the software.
[detailed version](doc/filegraph.svg)
Unfortunately, instagraal requires access to a GPU with CUDA drivers. It is currently not possible to make it compatible with singularity. This means instagraal has to be installed on the host machine for the scaffolding to work. It can also be run via docker by following instructions on the [official repository](https://github.com/koszullab/instaGRAAL).
### References
Tools used in this pipeline:
* filtlong v0.2.0:
* [flye](https://github.com/fenderglass/Flye/) v2.3.6: doi:10.1073/pnas.1604560113
* [seqtk](https://github.com/lh3/seqtk) v1.3
* [bowtie2](https://github.com/BenLangmead/bowtie2) v2.3.4.1: doi:10.1038/nmeth.1923
* [minimap2](https://github.com/lh3/minimap2) v2.17: doi:10.1093/bioinformatics/bty191
* [hypo](https://github.com/kensung-lab/hypo) v0.1.0: doi:10.1101/2019.12.19.882506
* [instagraal](https://github.com/koszullab/instaGRAAL) v0.1.6: doi:10.1038/ncomms6695
* [hicstuff](https://github.com/koszullab/hicstuff) v2.2.2: doi:10.5281/zenodo.2620608