BioNLP OST 2019 (AGAC Track)

The AGAC Task is part of the BioNLP Open Shared Tasks (BioNLP-OST: http://2019.bionlp-ost.org) and meets the BioNLP-OST standard of quality, originality and data formats.

Participants are welcomed to select tasks provided in AGAC Track. Among three subtasks, Task 1 is a traditional NER for 12 labels, which cultivate molecular phenomena related to gene mutation; Task 2 is a relation extraction task, which capture the thematic roles between entities; while Task 3 is a prediction task for the novel link discovery, which extract triple information among gene, function change, and disease out of the corpus texts.

The results of the AGAC Task will be presented at the BioNLP-OST workshop which is collocated with EMNLP-IJCNPL in Hong-Kong. Participating teams will be invited to submit their system description for publication in the proceedings of the workshop.

11 Mar, 2019. Sample data (50 texts) release.

10 Apr, 2019. Training data (250 texts) release.

12 Jul, 2019. Testing data (1000 texts) release.

12 Jul-19 Jul, 2019. Evaluation period.

1. We accept one round of results submission per day.
2. Please submit the result according to the "result submission template" in the Description page.
3. In the meantime, please submit a Project Abstract as well each time, which consists of title, authors, affiliations, and briefly describe the method.
4. Zip the result submission and project abstract, and submit it to xiajingbo.math@gmail.com.

19 Aug, 2019. Workshop paper submission due

TBD. Notification of paper acceptance

TBD. Camera ready paper submission

3 or 4 Nov, 2019. Workshop to be collocated with EMNLP-IJCNLP 2019 (Hong Kong)

28 Mar, 2020. Special issue submission due

(As a track in an open-shared task, AGAC track does not require any pre-registration.

Join Gmail group for contacting )

Identifying disease related genes and their mutations has long been a hot spot in biomedical community. One of the application is to help drug discovery by taking these genes as target genes. Therefore, the extraction of mutation-disease knowledge from PubMed is a worthy task. The gene-function change-disease knowledge in this track not only contains the relationship between mutation and disease, but also indicates the function change of the mutation, i.e., gain of function (GOF) and loss of function (LOF).

The purpose of AGAC track is to test the performance of various natural language processing (NLP) approaches on mutation-disease knowledge extraction in AGAC corpus.

• Sample Data (Released at 11 Mar 2019)

50 sample data. Link.

• Full training data ( Released at 10 Apr 2019)

250 texts with NER and Rel annotation labels for Task 1 and Task 2. (Click to download)

50 "Gene;Function change;disease" links for Task 3. (Click to download，submit format )

• Full testing data (Released at 12 Jul 2019)

1000 testing data. Link.

• Large-scale validation data

94 cancers related "Gene; Function Change; Disease" links for expanding the knowledge discovery based on Task 3, along with abstract level evidences. Link.

• Evaluation codes: Link(http://www.pubannotation.org/docs/evaluate-annotations/)

• Jingbo Xia, HZAU, xiajingbo.math@gmail.com
• Kaiyin Zhou, HZAU, kaiyinzhouhazu@gmail.com
• Yuxing Wang, HZAU, yuxingwang.www@gmail.com
• Mina Gachloo, HZAU, m_gachloo@yahoo.com
• AGAC Corpus Team, HZAU, http://xiajingbo.weebly.com/agac.html