Repository contains scripts to reproduce results in the paper as below:
Danang Crysnanto and Hubert Pausch. Sequence read mapping and variant discovery from bovine breed-specific augmented reference graphs. Biorxiv
Background
The bovine linear reference sequence was assembled from a DNA sample of a Hereford cow. It lacks diversity because it does not contain allelic variation. Lack of diversity is a drawback of linear references that results in reference allele bias. High nucleotide diversity and the separation of individuals by hundreds of breeds make cattle uniquely suited to investigate the optimal composition of variation-aware references.
Results
We augment the bovine linear reference sequence (ARS-UCD1.2) with variants filtered for allele frequency in dairy (Brown Swiss, Holstein) and dual-purpose (Fleckvieh, Original Braunvieh) cattle breeds to construct breed-specific and pan-genome reference graphs using the vg toolkit. We find that read mapping is more accurate to variation-aware than linear references if pre-selected variants are used for graph construction. Graphs that contain random variants don’t improve read mapping over the linear reference sequence. Breed-specific augmented graphs and pan-genome graphs enable almost similar accuracy improvements over the linear reference. We construct a whole-genome graph that contains the Hereford-based reference sequence and 14 million variants filtered for allele frequency in the Brown Swiss cattle breed. We show that our novel variation-aware reference facilitates accurate read mapping and unbiased sequence variant genotyping.
Conclusions
We developed the first variation-aware reference graph for an agricultural animal using variants that were filtered for allele frequency: https://doi.org/10.5281/zenodo.3570312. The novel reference structure improves sequence read mapping and variant genotyping over the linear reference. Our work may serve as a guideline to establish variation-aware reference structures in species with high genetic diversity and many sub-populations.
The paper contains four main parts, please go to respective pages for more details:
Part1: Variant prioritization
Part2 : Breeds graphs
Part3: Consensus genome
Part4: Variant genotyping
Note:
The data analyses utilized the ETH Zurich Leonhard Open High Performance Computing because of the high computing resources requirement. Reproducing in a local (dekstop) machine will not be possible in terms of memory and computing time.
However, final results are available in result folder and we have setup integration with binder, final data analyses can be repeated using launch binder button as above (also possible in local dekstop after cloning the repo)
The accompanying raw data for analyses are available via Zenodo, please download and untar-unzip the files. All raw data are available in data folder after unzipping.
tar -zxvf data.tar.gz
Danang Crysnanto
Animal Genomics ETH Zurich
Email: [email protected]
License: MIT
