BioCreative VI workshop, 2017 - Sérgio Matos
From IEETA
| Date | 2017/10/18 |
|---|---|
| Title | Identifying Relevant Literature for Precision Medicine Using Deep Neural Networks |
| Speaker | Sérgio Matos |
| Event | BioCreative VI workshop |
| Location | Washington DC |
| Country | USA |
| URL | http://www.biocreative.org/events/biocreative-vi/workshop/ |
Abstract: The aim of precision medicine is to select the best treatments for different patient groups, considering individual variability in genes, environment, and lifestyle. Regarding genetic variability, valuable information about variants and how they affect protein-protein interactions is available in the scientific literature. Extracting and curating this information in an efficient manner requires the application of text mining algorithms. In this paper, we describe deep learning classifiers for selecting, from the literature, documents containing information regarding the impact of genetic mutations on protein-protein interactions. Cross-validation results in the BioCreative VI Precision Medicine task corpus reached a precision of 0,698, recall of 0,735 and f-score of 0,715.