CoNECo: A corpus for named entity recognition and normalization of protein complexes

Data associated with the CoNECo corpus

The tagger software for which the dictionary files in tagger_dictionary_complex.tar.gz can be used with, can be found here: https://github.com/larsjuhljensen/tagger

The online version of the annotation documentation can be found here: https://katnastou.github.io/annodoc-CoNECo/ and a copy is here: Annodoc_CoNECo.pdf

The CoNECo corpus split in training, development and test sets in BRAT format can be found in CoNECo_corpus.tar.gz and in conll format: CoNECo_corpus_conll.tar.gz

The tagging results of Jensenlab tagger for the CoNECo test set are here: CoNECo_test_Jensenlab_tagger.tar.gz

The tagging results of Transformer-based tagger for the CoNECo test set are here: CoNECo_test_Transformer_tagger.tar.gz

The result from the large-scale run in entire PubMed (as of January 2024) and PMC Open Access (as of November 2023) articles for Jensenlab tagger is provided here: Jensenlab_tagger_large_scale_matches.tar.gz

The model used for the large scale run of the transformer-based method is here: CoNECo_Transformer_based_tagger_large_scale_model.tar.gz and the results from the large-scale tagging of the literature here: Transformer_based_tagger_large_scale_matches.tsv.gz

The counts and names of the common matches between the two large scale runs can be found here, while the counts for jensenlab tagger large scale unique matches and transformer-based tagger unique matches can be found here and here, respectively.