Microbial Genomics · AllTheBacteria
ATB
A standardised dataset of bacterial coding and intergenic sequences from approximately two million genomes.
What ATB is
ATB is a microbial genomics resource produced by AllTheBacteria. It provides a standardised collection of genomic elements drawn from approximately two million bacterial genome assemblies gathered across public sequence databases.
The dataset specifically details the genomic coordinates and sequences of both gene coding sequences and intergenic regions—the non-coding sequences located between genes—derived from high-quality assemblies. It is designed to support large-scale computational and comparative genomics across the bacterial domain.
How it was made
Bacterial genome assemblies are aggregated from public sequence repositories and standardised for comparative analysis. Sequences and coordinate positions corresponding to gene coding regions and non-coding intergenic spans are then extracted from the high-quality samples and structured into tabular Parquet files.
Key facts
- Publisher
- AllTheBacteria
- Category
- Microbial Genomics
- Assembly coverage
- Approximately 2 million bacterial assemblies
- Data format
- Parquet
- Modalities
- Tabular, text
- Size category
- 10B to 100B elements
What you can use it for
- Large-Scale Comparative Genomics
- Researchers can analyse gene coding sequences across millions of standardised bacterial assemblies to evaluate sequence conservation and gene distribution.
- Intergenic Region Characterisation
- Biologists can systematically inspect the positions and sequence structures of bacterial non-coding intergenic regions at scale.
- Microbial Machine Learning
- Data scientists can train sequence-based biological models using tabular sequence features readily loaded via tools such as Dask and Polars.
Explore the data
The schema, real rows, full-text search and a SQL filter — read in your browser, nothing downloaded.
Explore
Read the real rows without downloading anything
Reading rows…
LiveRead from AllTheBacteria/ATB at the moment you asked. Nothing is cached or stored — every row above came from that request.
How the columns are distributed
Over 73,331,458 rows of default/train
contig
string_text- min
- 24
- median
- 24
- mean
- 24.05
- max
- 26
end
int- min
- 1
- median
- 42,309
- mean
- 85525.66
- max
- 2.7M
feature_id
int- min
- 1
- median
- 74
- mean
- 147.50
- max
- 4,829
molecule
string_label- AA54%
- DNA46%
sample
string_text- min
- 12
- median
- 12
- mean
- 12.05
- max
- 14
sequence
string_text- min
- 1
- median
- 182
- mean
- 234.34
- max
- 13,280
source
string_label- ATB_0.2.incr.202408100%
start
int- min
- 1
- median
- 41,726
- mean
- 84963.44
- max
- 2.7M
strand
string_label- +73%
- -27%
Strengths and limits
Good at
- Integrates standardised sequence data derived from roughly two million bacterial genome assemblies.
- Isolates both coding sequences and intergenic regions specifically from high-quality assembly samples.
- Provided in columnar Parquet format to support high-throughput data processing workflows.
Watch out for
- Features are filtered to include only assemblies designated as high-quality samples rather than the complete repository.
- The data content is restricted to bacterial species and excludes archaea and other microbial groups.
Questions
Can I look inside ATB without downloading it?
Yes. This page reads its schema, real rows and per-column statistics in the browser, with full-text search and a SQL filter. Nothing is downloaded to your machine.
What licence is ATB under?
Its authors declare mit. Read the terms on the original page before reusing the data.
What sequence elements are catalogued in the dataset?
The dataset contains the positions and nucleotide sequences of both gene coding sequences and intergenic spaces.
How many genomes are represented in the AllTheBacteria collection?
The core AllTheBacteria resource comprises approximately two million bacterial genome assemblies aggregated from public sequence databases.
Which software libraries support working with this dataset?
The tabular Parquet files are structured for use with libraries including Datasets, Dask, Polars, and mlcroissant.
Written from ATB's own dataset card. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.