polars-bio—fast, scalable, and out-of-core operations on large genomic interval datasets
Bioinformatics · 2025
DOI: 10.1093/bioinformatics/btaf640 ↗Kurz gefasst
Fachartikel in der Zeitschrift Bioinformatics, 2025. Verfasst von Marek Wiewiórka und 3 weiteren Autor:innen. Beteiligte Einrichtung: Institute of Computer Science, Warsaw University of Technology, 00-665 Warsaw, Poland. Hauptthemen laut MEDLINE: Software, Genomics, Computational Biology. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 13 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Motivation Genomic studies very often rely on computationally intensive analyses of relationships between features, which are typically represented as intervals along a 1D coordinate system (such as positions on a chromosome). In this context, the Python programming language is extensively used for manipulating and analyzing data stored in a tabular form of rows and columns, called a DataFrame. Pandas is the most widely used Python DataFrame package and has been criticized for inefficiencies and scalability issues, which its modern alternative—Polars—aims to address with a native backend written in the Rust programming language. Results polars-bio is a Python library that enables fast, parallel and out-of-core operations on large genomic interval datasets. Its main components are implemented in Rust, using the Apache DataFusion query engine and Apache Arrow for efficient data representation. It is compatible with Polars and Pandas DataFrame formats. In a real-world comparison (107 versus 1.2×106 intervals), our library runs overlap queries 6.5×, nearest queries 15.5×, count_overlaps queries 38×, and coverage queries 15× faster than Bioframe. On equally sized synthetic sets (107 versus 107), the corresponding speedups are 1.6×, 5.5×, 6×, and 6×. In streaming mode, on real and synthetic interval pairs, our implementation uses 90× and 15× less memory for overlap, 4.5× and 6.5× less for nearest, 60× and 12× less for count_overlaps, and 34× and 7× less for coverage than Bioframe. Multi-threaded benchmarks show good scalability characteristics. To the best of our knowledge, polars-bio is the most efficient single-node library for genomic interval DataFrames in Python. Availability and implementation polars-bio is an open-source Python package distributed under the Apache License available for major platforms, including Linux, macOS, and Windows in the PyPI registry. The online documentation is https://biodatageeks.org/polars-bio/ and the source code is available on GitHub: https://github.com/biodatageeks/polars-bio and Zenodo: https://doi.org/10.5281/zenodo.16374290. are available at Bioinformatics online.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Autor:innen und Institutionen
- Institute of Computer Science, Warsaw University of Technology, 00-665 Warsaw, Poland.
- Department of Computer Science, University of Oxford, OX1 3QD Oxford, United Kingdom.
ORCID-Profile: Marek Wiewiórka, Tomasz Gambin
Zitieren
APA
Wiewiórka, M., Khamutou, P., Zbysiński, M., & Gambin, T. (2025). polars-bio—fast, scalable, and out-of-core operations on large genomic interval datasets. Bioinformatics. https://doi.org/10.1093/bioinformatics/btaf640
Vancouver
Wiewiórka M, Khamutou P, Zbysiński M, Gambin T. polars-bio—fast, scalable, and out-of-core operations on large genomic interval datasets. Bioinformatics. 2025. doi:10.1093/bioinformatics/btaf640
BibTeX
@article{Wiewirka2025,
author = {Wiewiórka, Marek and Khamutou, Pavel and Zbysiński, Marek and Gambin, Tomasz},
title = {polars-bio—fast, scalable, and out-of-core operations on large genomic interval datasets},
journal = {Bioinformatics},
year = {2025},
doi = {10.1093/bioinformatics/btaf640}
}
EUVIMED