Published February 19, 2025
| Version v3
Dataset
Open
BioToFlow: a corpus annotated with bioinformatics workflows information
Authors/Creators
- 1. Université Paris-Saclay, CNRS, LISN, 91400, Orsay, France
- 2. Université Paris-Saclay, CEA, List, F-91120, Palaiseau, France
Description
BioToFlow is a corpus describing bioinformatics workflows in English publications. These annotations are available in the BRAT Rapid Annotation Tool (BRAT) standoff format (https://brat.nlplab.org/standoff.html).
This corpus is composed of 52 articles (26 articles related to Nextflow workflows and 26 on Snakemake workflows, randomly selected from PubMed) with a total of 78 419 tokens 27 786 annotated tokens.
Repository organisation
The articles are separated into two directories:
- one containing all the articles for the training phases (39) and
- the other with 13 articles for test.
Papers
Please cite BioToFlow in any research that uses or extends it :
- Sebe, C., Cohen-Boulakia, S., Ferret, O., Névéol, A.: Extracting information in a low-resource setting: Case study on bioinformatics workflows (2024), https://arxiv.org/abs/2411.19295
In this article accepted to IDA 2025 (in English), we present *BioToFlow* and experiments with few shot named entity recognition (NER) using an autoregressive language model, we also use a pre-existing corpus and test integration of workflow knowledge in NER models.
- Clémence Sebe, Sarah Cohen-Boulakia, Olivier Ferret, Aurélie Névéol. Extraction d’entités nommées décrivant des chaînes de traitement bioinformatiques dans des articles scientifiques en anglais. 35emes Journées d’Études sur la Parole (JEP 2024) 31eme Conférence sur le Traitement Automatique des Langues Naturelles (TALN 2024) 26eme Rencontre des Étudiants Chercheurs en Informatique pour le Traitement Automatique des Langues (RECITAL 2024), Jul 2024, Toulouse, France. pp.422-434. hal-04623033.
In this article (in French), we present the second version of *BioToFlow*. The new articles are annotated with entities and attributes. We conduct preliminary experiments with NER with a specific focus on the memorization vs. generalization abilities of statistical and rule-based methods.
- Sebe C., Névéol A., Cohen-Boulakia S. & Gaignard A. (2023). Extraction d’informations sur les workflows scientifiques à partir de la littérature. volume Extraction et Gestion des Connaissances, RNTI-E-39, p. 313.
In this article (in French), we present the first version of the corpus with 24 articles annotated with entities and relations. We show the feasibility of the task of NER.
Contact
- Clémence Sebe, clemence.sebe@universite-paris-saclay.fr
Funding
This work received support from the National Research Agency under the France 2030 program, with reference to ANR-22-PESN-0007.Files
BioToFlow.zip
Files
(508.3 kB)
| Name | Size | Download all |
|---|---|---|
|
md5:6dee2e44b3476906c01680c58df83fbc
|
508.3 kB | Preview Download |