Text Mining for Drug–Drug Interaction
In order to understand the mechanisms of drug-drug interaction (DDI), the study of pharmacokinetics (PK), pharmacodynamics (PD), and pharmacogenetics (PG) data are significant. In recent years, drug PK parameters, drug interaction parameters, and PG data have been unevenly collected in different databases and published extensively in literature. Also the lack of an appropriate PK ontology and a well-annotated PK corpus, which provide the background knowledge and the criteria of determining DDI,
... respectively, lead to the difficulty of developing DDI text mining tools for PK data collection from the literature and data integration from multiple databases. To conquer the issues, we constructed a comprehensive pharmacokinetics ontology. It includes all aspects of in vitro pharmacokinetics experiments, in vivo pharmacokinetics studies, as well as drug metabolism and transportation enzymes. Using our pharmacokinetics ontology, a PK corpus was constructed to present four classes of pharmacokinetics abstracts: in vivo pharmacokinetics studies, in vivo pharmacogenetic studies, in vivo drug interaction studies, and in vitro drug interaction studies. A novel hierarchical three-level annotation scheme was proposed and implemented to tag key terms, drug interaction sentences, and drug interaction pairs. The utility of the pharmacokinetics ontology was demonstrated by annotating three pharmacokinetics studies; and the utility of the PK corpus was demonstrated by a drug interaction extraction text mining analysis. The pharmacokinetics ontology annotates both in vitro pharmacokinetics experiments and in vivo pharmacokinetics studies. The PK corpus is a highly valuable resource for the text mining of pharmacokinetics parameters and drug interactions. There are two basic types of drug interaction, pharmacokinetics (PK) and pharmacodynamics (PD). In short, PK investigates the activity of drug combinations with drug absorption, disposition, metabolism, excretion, and transportation (ADMET), which describes how these five criteria influence drug level (concentration). Pharmacokinetically speaking, potentiative or reductive combinations are, respectively, correlated to positive or negative modulation of drug transport, permeation, distribution, localization, or metabolism. Potentiative modulation of drug transport will enhance drug absorption via the disruption of transport carrier, increase drug concentration in plasma by inhibiting metabolic process, and stimulate or inhibit the metabolism of drugs into active or inactive form. On the other hand, reductive modulation provides contrasting perspectives to potentiative modulation. The reductive modulation of drug transport typically blocks drug absorption, decreases drug concentration in plasma, and reduces drug metabolism activity  . Those information brings to systematically investigate the physiological and biochemical mechanisms of drug exposure in multiple tissue types, cells, animals, and human subjects  , which links preclinical and clinical phase of drug development. If the PK can be interpreted as the doseconcentration relationship, pharmacodynamics (PD) can be defined as the mechanism of drug action and relationship between drug concentration and effect. A drug's pharmacodynamics effect ranges widely from the molecular signals (such as its targets or downstream biomarkers) to clinical symptoms (such as the efficacy or side effect endpoints). Classification of its therapeutic effects: It can be synergistic, additive, or antagonistic if the effect is greater than, equal to, or less than the summed effects of drug combinations  . Wu et al. Biomedical Text Mining-Text mining refers to the process of deriving highquality information from text, which relies on NLP. To translate the text into computerreadable language, there are some basic steps of NLP , including sentence splitting, tokenization, part of speech, named entity recognition (NER), shallow parsing, and syntactic parsing. In this section, we do not go into the details of techniques for NLP tools. The attentions will be paid more on the tasks of corpus construction, IR, or information extraction (IE), which employs highly scalable statistics-based techniques to index and search large volume of text efficiently. Extracting facts from texts is the goal of text-mining systems. The range of extraction tasks can be narrow from retrieving potentially relevant articles by sophisticated keyword search Wu et al.