Table 2 Relative strengths of five pangenome graph construction tools. Explanation of rows: (1) efficacy of construction algorithm, measuring wall-clock time; (2) degree to which variants (e.g., SNPs) are retained; (3) ability of a tool to perform well on large datasets, both in comparison to other tools but also compared to smaller datasets; (4) ability to modify the produced data structure to add or remove haplotypes; (5) property of producing the same result irrespective of perturbations, such as permutation of the input order, and repeating the same run; (6) existence of tools (and operations) that can be applied to the resulting graphs; (7) whether input haplotypes information is retained by the tools, and if so, its space efficiency; (8) whether the entire graph can be directly visualized and interpreted; (9) easiness of “zooming in” a specific genomic region and interpret variants; (10) summarizes the functionalities provided by the tools to annotate the pangenomes with genomic data; (11) ability to share information between the graph and a linear reference
From: Comparing methods for constructing and representing human pangenome graphs
Metric | Bifrost | pggb | Minigraph-Cactus | Minigraph | mdbg |
|---|---|---|---|---|---|
1) Construction speed | \(\bullet \bullet \circ\) | \(\bullet \circ \circ\) | \(\bullet \circ \circ\) | \(\bullet \bullet \circ\) | \(\bullet \bullet \bullet\) |
2) Variations | \(\bullet \bullet \bullet\) | \(\bullet \bullet \bullet\) | \(\bullet \bullet \bullet\) | \(\bullet \bullet \circ\) | \(\bullet \bullet \circ\) |
3) Scalablilty | \(\bullet \bullet \bullet\) | \(\bullet \circ \circ\) | \(\bullet \circ \circ\) | \(\bullet \bullet \circ\) | \(\bullet \bullet \bullet\) |
4) Editability | \(\bullet \bullet \bullet\) | \(\bullet \bullet \circ\) | \(\bullet \circ \circ\) | \(\bullet \bullet \circ\) | \(\bullet \circ \circ\) |
5) Stability | \(\bullet \bullet \bullet\) | \(\bullet \circ \circ\) | \(\bullet \circ \circ\) | \(\bullet \bullet \circ\) | \(\bullet \bullet \bullet\) |
6) Accessibility by downstream applications | \(\bullet \circ \circ\) | \(\bullet \bullet \bullet\) | \(\bullet \bullet \bullet\) | \(\bullet \bullet \circ\) | \(\bullet \circ \circ\) |
7) Haplotype compression performance | \(\bullet \bullet \circ\) | \(\bullet \bullet \bullet\) | \(\bullet \bullet \bullet\) | \(\bullet \circ \circ\) | \(\bullet \circ \circ\) |
8) Ease of visualization | \(\bullet \circ \circ\) | \(\bullet \bullet \circ\) | \(\bullet \bullet \circ\) | \(\bullet \bullet \bullet\) | \(\bullet \bullet \bullet\) |
9) Loci visualization and interpretability | \(\bullet \circ \circ\) | \(\bullet \bullet \circ\) | \(\bullet \bullet \bullet\) | \(\bullet \bullet \circ\) | \(\bullet \circ \circ\) |
10) Metadata and annotation | \(\bullet \bullet \circ\) | \(\bullet \bullet \bullet\) | \(\bullet \bullet \circ\) | \(\bullet \circ \circ\) | \(\bullet \circ \circ\) |
11) Compatibility with a linear reference coordinates | \(\bullet \circ \circ\) | \(\bullet \bullet \bullet\) | \(\bullet \bullet \bullet\) | \(\bullet \bullet \circ\) | \(\bullet \circ \circ\) |