Accessing our in-house literature mining tool for your projects saves a lot of time and gives added value to the project. The platform can be used to query large biological or chemical datasets for a given project, and is also able to find associations that might not be present in manually curated databases.
Unique
- Updated weekly to accommodate new data as soon as it is available.
- Built on top of industry standard Solr/Lucene engine with query response times of less than 15 milliseconds for complex queries.
- Drug-Target-Disease associations
- 90,000 important biomedical compounds.
- 4000 diseases within a custom curated ontology, including custom developed rare disease knowledgebase.
- 20,400 protein targets
Below mentioned are some of the areas we have utilized the potential of our literature mining tool and it has played a vital role.
Google's Bag-of-Words model used to extract synonyms
- Given a word, train a neural network to recognize the "context" of the word.
- Given the context, recognize the word.
- Given word, reproduce the context.
- For instance, "TB" and "Tuberculosis" may have similar contexts in the sentences.
- TB kills 2 million people every year.
- Tuberculosis related deaths have reached 2 million according to WHO estimates.
Semantic relationships using deep learning methods
- Given multiple instances of a type of relationship, find others of the same type.
- TREATMENT type
A cures B or A is a treatment for B
- CAUSE type
A causes B or A leads to B
- Based on word2vec models.
- Trained deep learning models for 8 different types of relationships.
- High accuracy.
To explore our services, write to us at info@excelra.com