Datasets and code used in "Theseus: Fast and Optimal Affine-Gap Sequence-to-Graph Alignment"

Datasets used in the "Theseus: Fast and Optimal Affine-Gap Sequence-to-Graph Alignment" paper. The data is compressed into the theseus_datasets.tar.gz file. This file includes the datasets for the two experiments on the paper:MSA datasets: File Size Source hiv_100.fasta 100 sequences, each of approximately 10Kbp https://github.com/niemasd/ViralMSA/blob/master/example/example_hiv.fas mtb_benchmark_50kbp_shortened.fna 342 sequences, each of approximately 50Kbp Derived from a set of 342 RefSeq-complete whole-genome assemblies of Mycobacterium tuberculosis genomes. Sequences start at the dnaA gene and are 50Kbp long. mtb_benchmark_250kbp_trmB.fna 342 sequences, each of approximately 250Kbp Derived from a set of 342 RefSeq-complete whole-genome assemblies of Mycobacterium tuberculosis genomes. Sequences start at the dnaA gene and are truncated at the trmB gene. mtb_benchmark_500kbp_thiE.fna 342 sequences, each of approximately 500Kbp Derived from a set of 342 RefSeq-complete whole-genome assemblies of Mycobacterium tuberculosis genomes. Sequences start at the dnaA gene and are truncated at the thiE gene. mtb_benchmark_1Mbp_gltA2.fna 342 sequences, each of approximately 1Mbp Derived from a set of 342 RefSeq-complete whole-genome assemblies of Mycobacterium tuberculosis genomes. Sequences start at the dnaA gene and are truncated at the gltA2 gene. covid_19_complete.fasta 2732 sequences, each of approximately ∼30 Kbpin length 2732 GenBank-complete SARS-CoV-2 genome assemblies. monkeypox_100_seq.fasta 100 sequences, each of approximately ∼200 Kbp in length 100 RefSeq-complete whole genome assemblies of Monkey pox’s virus. Sequence-to-graph/pangenome read mapping datasets: File Size Source SRR062634_1.filt_REDUCED.fasta 250K sequences of length 100bp Human Pangenome Reference Consortium 211109_M024_V350038332_L01_HUMuarfR092940-606_1_REDUCED.fasta 250K sequences of length 150bp NIST Genome in a Bottle (GIAB) project D1_S1_L001_R1_001_REDUCED.fasta 250K sequences of length 250bp NIST Genome in a Bottle (GIAB) project This dataset is derived from original data produced by third parties, as detailed above. All rights to the original data remain with the original authors or copyright holders. Users are responsible for ensuring compliance with the licensing terms of the original data sources. Sequence-to-cyclic-graph datasets: These datasets have been synthetically generated to test the ability of Theseus to align against cyclic reference graphs, comparing it to the graph unfolding strategy used by vg map. The synthetic cyclic graphs with a controlled structure. For that, we create backbone graphs with N in {10, 100, 1000} nodes, with node sequences of average length 50 bp, and connect each node to its successor to ensure graph connectivity. Then, we add a bounded number of random edges per node. These random edges can connect to nearby forward nodes or to previously created nodes, introducing cycles into the graph. For each graph size, we generate queries of length 100, 150, and 250 bp. We have nine pairs of reference graph and queries. This happens because we have three graph sizes, with N in {10, 100, 1000} nodes, and 3 query sizes, of 100, 150, and 250 base pairs. Code snapshot: This dataset also contains a snapshot of the code use to conduct the experiments in the manuscript, corresponding to Theseus v0.1 on Github.