GraphChainer

GraphChainer is an accurate aligner of long reads to a variation graph, based on co-linear chaining.

Compiling

To compile, run these:

• Install miniconda
• git submodule update --init --recursive
• conda env create -f CondaEnvironment.yml
• conda activate GraphChainer
• make bin/GraphChainer

Running

Quickstart: ./bin/GraphChainer -t 4 -f reads.fastq -g graph.gfa -a out.gam

Key parameters:

• -t Number of threads (optional, default 1).
• -f Input reads. Format .fasta / .fastq / .fasta.gz / .fastq.gz. You can input multiple files with -f file1 -f file2 ... or -f file1 file2 ....
• -g Input graph, format .gfa / .vg. This graph must be acyclic, see below how to construct an acyclic graph with vg.
• -a Output file name. Format .gam or .json.

Parameters related to colinear chaining:

• --sampling-step <double> Sampling step factor (default 1). Use >1 (<1, >0) for faster (slower), but less (more) accurate alignments. It increases (decreases) the sampling sparsity of fragments.
• --colinear-split-len <int> The length of the fragments in which the long read is split to create anchors (default 35).
• --colinear-split-gap <int> The distance between consecutive fragments (default 35). If --sampling-step is set, then always --colinear-split-gap = ceil(--sampling-step * --colinear-split-len).
• --colinear-gap <int> When converting an optimal chain of anchors into an alignment path, split the path if the distance in the graph between consecutive anchors is greater than this value (default 10000).

Constructing an (acyclic) variation graph

Use vg and run:

vg construct -t 30 -a -r {ref} -v {vcf} -R 22 -p -m 3000000

Datasets availability

The graphs built for the experiments of GraphChainer can be found in Zenodo at https://doi.org/10.5281/zenodo.7729494 , https://doi.org/10.5281/zenodo.6875064 and at https://doi.org/10.5281/zenodo.6587252

The real read sets can be found in Zenodo ar TODO

The evaluation pipeline used in the paper can be found at https://github.com/algbio/GraphChainer-scripts

Citation

If you use GraphChainer, please cite as:

Jun Ma, Manuel Cáceres, Leena Salmela, Veli Mäkinen, Alexandru I. Tomescu. Chaining for Accurate Alignment of Erroneous Long Reads to Acyclic Variation Graphs. Submitted, 2022

Credits

GraphChainer is built on the excellent code base of GraphAligner, which is released under MIT License. GraphAligner is described in the paper GraphAligner: Rapid and Versatile Sequence-to-Graph Alignment by Mikko Rautiainen and Tobias Marschall.