---
task_categories:
- text-classification
- token-classification
language:
- en
tags:
- biology
- nlp
- ner
pretty_name: CellFinder BRAT
size_categories:
- 1K<n<10K
---

# Dataset Card for CellFinder BRAT

<!-- Provide a quick summary of the dataset. -->

This data is a direct download of a Humboldt university of Berlin resource.

## Dataset Details

### Dataset Description

<!-- Provide a longer summary of what this dataset is. -->

From data source _https://www.informatik.hu-berlin.de/de/forschung/gebiete/wbi/resources/cellfinder/_, this data is described as follows:

**- - - - - - -**

The first version of our corpus is composed of 10 full text documents containing more than 2,100 sentences, 65,000 tokens and 5,200 annotations for entities.
The corpus has been annotated with six types of entities (_anatomical parts, cell components, cell lines, cell types, genes/protein and species_) with an overall inter-annotator agreement around 80%.

**- - - - - - -**

- **Curated by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]

### Dataset Sources

<!-- Provide the basic links for the dataset. -->

- **Link:** _https://www.informatik.hu-berlin.de/de/forschung/gebiete/wbi/resources/cellfinder/_
- **Paper:** Mariana Neves, Alexander Damaschun, Andreas Kurtz, Ulf Leser. Annotating and evaluating text for stem cell research. Third Workshop on Building and Evaluation Resources for Biomedical Text Mining (BioTxtM 2012) at Language Resources and Evaluation (LREC) 2012.

## Uses

<!-- Address questions around how the dataset is intended to be used. -->
We utilised this dataset to train a NER model to annotate the entity types detailed by the authors of this dataset.

In turn this data became part of a larger dataset for our group.

We copied this format of the data to our huggingface project for easier ingestion into our NER-model production pipeline. 


## Dataset Structure

<!-- This section provides a description of the dataset fields, and additional information about the dataset structure such as criteria used to create the splits, relationships between data points, etc. -->

This data is stored in the BRAT format, with each article, identified through it's PMCID, annotated via a _.txt_ and a _.ann_ file.

Annotations within the .ann file appear as shown in the following example:

**- - - - - - -**

T312	CellType 13239 13246	T cells

T313	GeneProtein 13235 13238	CD4

T314	CellType 13185 13196	macrophages

T315	Species 16981 16986	HIV-1

**- - - - - - -**

With columns corresponding to: _ID    Entity type    Start    End    Phrase_

### Recommendations

<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->

Users should be made aware of the risks, biases and limitations of the dataset. More information needed for further recommendations.

## Citation

Mariana Neves, Alexander Damaschun, Andreas Kurtz, Ulf Leser. Annotating and evaluating text for stem cell research. Third Workshop on Building and Evaluation Resources for Biomedical Text Mining (BioTxtM 2012) at Language Resources and Evaluation (LREC) 2012.


## Dataset Card Authors

Christine Withers - _https://huggingface.co/christine-withers_

## Dataset Card Contact

Christine Withers - _https://huggingface.co/christine-withers_