pyexpressionatlas

A Python client for searching and downloading gene expression datasets from EMBL-EBI Expression Atlas, providing full compatibility with the R Bioconductor package.

[!NOTE] This package is a fork of expression-atlas to use BiocPy data structures.

Install

To get started, install the package from PyPI

pip install pyexpressionatlas

Get Started

Expression Atlas is a comprehensive resource of gene and protein expression data across species and biological conditions. This Python package provides programmatic access to:

• Search: Query thousands of curated RNA-seq and microarray experiments
• Download: Retrieve experiment data with automatic format handling
• Analyze: Work with R-compatible data structures in Python

Basic Usage

from pyexpressionatlas import ExpressionAtlasClient

# Initialize client (optionally specify a custom cache directory)
client = ExpressionAtlasClient(cache_dir="~/.cache/my_custom_cache")

# Search for experiments
results = client.search_experiments(
    properties=["cancer", "breast"],
    species="homo sapiens"
)
print(results)

BiocFrame with 208 rows and 4 columns
        Accession      Species                    Type                   Title
            <list>       <list>                  <list>                  <list>
[0]  E-MTAB-8198          None                    None Functional effect of...
[1]  E-MTAB-8532          None                    None DNA microarray studi...
[2] E-GEOD-43306          None                    None Translating transcri...
            ...          ...                     ...                     ...
[205]   E-MTAB-779        None                    None OncomiRs like let-7 ...
[206]  E-TABM-1118        None                    None Transcrption profili...
[207]   E-TABM-601        None                    None Transcription profil...

Fetch Full Metadata

The initial search is optimized for speed and does not fetch full metadata. To retrieve complete details (including Species and Type), use fetch_experiment_metadata:

# Fetch full metadata for specific experiments
metadata = client.fetch_experiment_metadata(["E-MTAB-8198", "E-MTAB-8532"])
print(metadata)

BiocFrame with 2 rows and 4 columns
        Accession      Species                    Type                   Title
           <list>       <list>                  <list>                  <list>
[0]   E-MTAB-8198 Homo sapiens Cell line - High-thr... Functional effect of...
[1]   E-MTAB-8532 Homo sapiens Human - One-color mi... DNA microarray studi...

Download RNA-seq Data

# Download a single experiment
exp = client.get_experiment("E-MTAB-1625")

# Access RNA-seq data (SummarizedExperiment)
rnaseq = exp["rnaseq"]
counts = rnaseq.assay("counts")  # numpy array: genes × samples

print(f"Shape: {counts.shape[0]} genes × {counts.shape[1]} samples")
# Shape: 58735 genes × 24 samples

# Sample metadata (BiocFrame)
sample_info = rnaseq.get_column_data()
print(sample_info.get_column_names())
# ['cell line', 'compound', 'developmental stage', 'disease', 'dose', 'genotype', 'organism', 'organism part']

# Gene annotations (BiocFrame)
gene_info = rnaseq.get_row_data()
print(gene_info.shape)
# (58735, 1)

print(rnaseq)

class: SummarizedExperiment
dimensions: (58735, 24)
assays(1): ['counts']
row_data columns(1): ['Gene Name']
row_names(58735): ['ENSG00000000003', 'ENSG00000000005', 'ENSG00000000419', ..., 'ENSG00000285992', 'ENSG00000285993', 'ENSG00000285994']
column_data columns(8): ['cell line', 'compound', 'developmental stage', 'disease', 'dose', 'genotype', 'organism', 'organism part']
column_names(24): ['ERR3456453', 'ERR3456442', 'ERR3456443', ..., 'ERR3456450', 'ERR3456459', 'ERR3456444']
metadata(2): accession source

Batch Downloads

# Download multiple experiments
accessions = results.get_column("Accession")[:10]
experiments = client.get_experiments(accessions)

# Access individual experiments
for acc, exp in experiments.items():
    if exp is not None:
        print(f"{acc}: {exp['rnaseq'].shape if 'rnaseq' in exp else 'microarray'}")

Caching Mechanism

To optimize performance and reduce load on the FTP servers, all data downloads are automatically cached locally using pyBiocFileCache.

• By default, the cache is stored at ~/.cache/expressionatlas_bfc.
• You can customize this location when initializing the client by passing the cache_dir argument: client = ExpressionAtlasClient(cache_dir="/path/to/custom/cache").

Direct RData / rda Support

The client automatically downloads and parses both .rds and .Rdata / .rda files directly without relying on a cloud converter service:

• Tries downloading and parsing .rds file using rds2py.read_rds.
• If the .rds file is not available, falls back to direct download and loading of the .Rdata file using rds2py.read_rda.

Note

This project has been set up using BiocSetup and PyScaffold.