Morphology analysis of unlabeled red blood cells based on quantitative differential phase contrast microscopy.

The current classical blood smear technique to observe the morphology of single red blood cells (RBCs) for classification is a laborious and error-prone process. To objectively evaluate the morphology of blood cells, we established a method of computational imaging based on a programmable light emitting diode array. By using quantitative differential phase contrast (qDPC), we characterized the morphology of unlabeled RBCs as well as blood smears. By focusing on comparing the difference of imaging between unlabeled RBCs and stained RBCs under multimode microscopic imaging technology, we demonstrated that qDPC could clearly differentiate discocytes and spherocytes in both unlabeled RBCs and blood smears. The phase map provided by quantitative phase imaging further enhanced the classification accuracy. According to statistical analysis from morphological indexes, the qDPC imaging has a significantly improvement in non-circularity, texture inhomogeneity and equivalent diameters of cells. Thus, this method has a significant superiority in the capability to analyze the morphology of RBCs and could be applied to clinical assays for determining morphological, functional, and structural deterioration of RBCs.

Drug name recognition in biomedical texts: a machine-learning-based method.

Currently, there is an urgent need to develop a technology for extracting drug information automatically from biomedical texts, and drug name recognition is an essential prerequisite for extracting drug information. This article presents a machine-learning-based approach to recognize drug names in biomedical texts. In this approach, a drug name dictionary is first constructed with the external resource of DrugBank and PubMed. Then a semi-supervised learning method, feature coupling generalization, is used to filter this dictionary. Finally, the dictionary look-up and the condition random field method are combined to recognize drug names. Experimental results show that our approach achieves an F-score of 92.54% on the test set of DDIExtraction2011.

Extracting drug-drug interaction from the biomedical literature using a stacked generalization-based approach.

Drug-drug interaction (DDI) detection is particularly important for patient safety. However, the amount of biomedical literature regarding drug interactions is increasing rapidly. Therefore, there is a need to develop an effective approach for the automatic extraction of DDI information from the biomedical literature. In this paper, we present a Stacked Generalization-based approach for automatic DDI extraction. The approach combines the feature-based, graph and tree kernels and, therefore, reduces the risk of missing important features. In addition, it introduces some domain knowledge based features (the keyword, semantic type, and DrugBank features) into the feature-based kernel, which contribute to the performance improvement. More specifically, the approach applies Stacked generalization to automatically learn the weights from the training data and assign them to three individual kernels to achieve a much better performance than each individual kernel. The experimental results show that our approach can achieve a better performance of 69.24% in F-score compared with other systems in the DDI Extraction 2011 challenge task.