macwiatrak/bacbench-strain-clustering-dna

Dataset for whole-bacterial genomes clustering (DNA)

A dataset of 60,710 bacterial genomes across 25 species, 10 genera and 7 families.

The genome DNA sequences have been extracted from MGnify. Each row contains whole bacterial genome, with spaces separating different contigs present in the genome.

Labels

The species, genus and family labels have been provided by MGnify

Usage

We recommend loading the dataset in a streaming mode to prevent memory errors.

from datasets import load_dataset


ds = load_dataset("macwiatrak/bacbench-strain-clustering-dna", split="test", streaming=True)

# fetch the DNA sequence of the genome
item = next(iter(ds))
dna_seq = item["dna_sequence"]
# get DNA sequences for each contig
dna_seq_contigs = dna_seq.split()

Split

The strain clustering task is a fully unsupervised task, therefore, we do not perform any split and the whole dataset is available as part of the test split.

See github repository for details on how to embed the dataset with DNA and protein language models as well as code to predict antibiotic resistance from sequence. For coding sequence representation of the genome see the strain-clustering-protein-sequences dataset.

---

dataset_info: features: - name: genome_name dtype: string - name: contig_name sequence: string - name: dna_sequence dtype: string - name: start sequence: sequence: string - name: end sequence: sequence: string - name: locus_tag sequence: sequence: string - name: strand sequence: sequence: string - name: genome_completeness dtype: string - name: genome_lineage dtype: string - name: genome_sample_accession dtype: string - name: genome_study_accession dtype: string - name: country dtype: string - name: family dtype: string - name: genus dtype: string - name: species dtype: string splits: - name: test num_bytes: 169603550566 num_examples: 60710 download_size: 78407309865 dataset_size: 169603550566 configs: - config_name: default data_files: - split: test path: data/test-*