CeLLaTe

A resolved representation of the CeLLaTe Dataset Overview We present a dataset for Cell Type, Cell Line, and Tissue entities, named CeLLaTe (CeLl type, Cell Line and Tissue entities). The resulting dataset comprises 34 full-text articles, 7,800 sentences, and 8,113 annotated entity spans, with 40% of sentences containing at least one annotation. Annotation guidelines were developed to ensure consistency, producing data with an inter-annotator agreement score (Cohen's Kappa coefficient) of 0.925. This dataset represents an un-edited, whole version of CeLLaTe. There are no splits, and no entity tags outside of those from the base dataset schema: CellType CellLine Tissue This version of CeLLaTe allows for a view which is not manipulated or obscured by any downstream tasks carried out in efforts to train transformer models. On 'vague' terms Two versions of this dataset have been created: one retaining 'vague' entities and one excluding them (this dataset). 'Vague' terms were removed through the use of a vague terms dictionary, seen at: https://github.com/EuropePMC/OTAR3088/blob/main/docs/vague_entity_examples.tsv The filtering of 'vague' entities applies exclusively to the subsects of CeLLaTe which were manually curated. These are the splits which contain articles of 'single-cell transcriptomics' and 'pharmacology' themes. We define vague entities as terms that do not strictly satisfy the criteria of a well-defined named entity, but may exhibit entity-like characteristics depending on contextual usage. Dataset Schema Columns are as follows: sentence: a single sentence extracted from a biomedical article entities: a list of entity annotations associated with the sentence data_source: the originating corpus or article collection from which the sentence was derived Annotations are provided at the sentence level to facilitate downstream NER training, evaluation, and AL-driven re-annotation workflows Data Sources and Domain Composition The dataset integrates articles from three complementary biomedical domains, each contributing distinct entity distributions: 1. Single cell transcriptomics Literature: High prevalence of CellType and Tissue entities Rich terminology diversity Manually curated 2. Pharmacology literature Enriched in CellLine mentions MedChem articles were retrieved here from the ChEMBL 34 database SQLite queries linked the ASSAY DESCRIPTIONS and DOCS tables, collecting supporting literature for assay descriptions mentioning cell lines Manually curated 3. Stem Cell Research (CellFinder) Contains all three entity types Particularly rich in CellType mentions Historically curated dataset with expert annotations (Dated more than 10 years ago) This multi-domain composition allows the evaluation of: Cross-domain robustness Entity distribution shifts Label imbalance behaviour Domain adaptation strategies Stem Cell Article Source (CellFinder) Stem cell–related articles were obtained from the CellFinder repository.The original dataset and annotation methodology are described in: Mariana Neves, Alexander Damaschun, Andreas Kurtz, Ulf Leser (2012) Annotating and evaluating text for stem cell research. In Proceedings Third Workshop on Building and Evaluation Resources for Biomedical Text Mining (BioTxtM 2012), Language Resources and Evaluation (LREC) 2012. The CellFinder corpus provides historically curated annotations across multiple stem-cell–related entity types.