Non-small cell lung cancer (NSCLC) constitutes the most common type of lung cancer and is frequently diagnosed at advanced stages. Clinical studies have shown that molecular targeted therapies increase survival and improve quality of life in patients. Nevertheless, the realization of personalized therapies for NSCLC faces a number of challenges including the integration of clinical and genetic data and a lack of clinical decision support tools to assist physicians with patient selection. To address this problem, we used frequent pattern mining to establish the relationships of patient characteristics and tumor response in advanced NSCLC. Univariate analysis determined that smoking status, histology, epidermal growth factor receptor (EGFR) mutation, and targeted drug were significantly associated with response to targeted therapy. We applied four classifiers to predict treatment outcome from EGFR tyrosine kinase inhibitors. Overall, the highest classification accuracy was 76.56% and the area under the curve was 0.76. The decision tree used a combination of EGFR mutations, histology, and smoking status to predict tumor response and the output was both easily understandable and in keeping with current knowledge. Our findings suggest that support vector machines and decision trees are a promising approach for clinical decision support in the patient selection for targeted therapy in advanced NSCLC.

An important task in biological information extraction is to identify descriptions of biological relations and events involving genes or proteins. In this work, we propose a graph‐based approach to automatically learn rules for detecting biological events in the life science literature. The event rules are learned by identifying the key contextual dependencies from full parsing of annotated text. The detection is performed by searching for isomorphism between event rules and the dependency graphs of complete sentences. When applying our approach to the data sets of the Task 1 of the BioNLP‐ST 2009, we achieved a 40.71%F‐score in detecting biological events across nine event types. Our 56.32% precision is comparable with the state‐of‐the‐art systems. The approach may also be generalized to extract events from other domains where training data are available because it requires neither manual intervention nor external domain‐specific resources. The subgraph matching algorithm we developed is released under the new BSD license and can be downloaded from http://esmalgorithm.sourceforge.net .

MOTIVATION: Aligning protein sequences with the best possible accuracy requires sophisticated algorithms. Since the optimal alignment is not guaranteed to be the correct one, it is expected that even the best alignment will contain sites that do not respect the assumption of positional homology. Because formulating rules to identify these sites is difficult, it is common practice to manually remove them. Although considered necessary in some cases, manual editing is time consuming and not reproducible. We present here an automated editing method based on the classification of 'valid' and 'invalid' sites. RESULTS: A support vector machine (SVM) classifier is trained to reproduce the decisions made during manual editing with an accuracy of 95.0%. This implies that manual editing can be made reproducible and applied to large-scale analyses. We further demonstrate that it is possible to retrain/extend the training of the classifier by providing examples of multiple sequence alignment (MSA) annotation. Near optimal training can be achieved with only 1000 annotated sites, or roughly three samples of protein sequence alignments. AVAILABILITY: This method is implemented in the software MANUEL, licensed under the GPL. A web-based application for single and batch job is available at http://fester.cs.dal.ca/manuel. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

BACKGROUND: In protein evolution, the mechanism of the emergence of novel protein domain is still an open question. The incremental growth of protein variable regions, which was produced by stochastic insertions, has the potential to generate large and complex sub-structures. In this study, a deterministic methodology is proposed to reconstruct phylogenies from protein structures, and to infer insertion events in protein evolution. The analysis was performed on a broad range of SCOP domain families. RESULTS: Phylogenies were reconstructed from protein 3D structural data. The phylogenetic trees were used to infer ancestral structures with a consensus method. From these ancestral reconstructions, 42.7% of the observed insertions are nested insertions, which locate in previous insert regions. The average size of inserts tends to increase with the insert rank or total number of insertions in the variable regions. We found that the structures of some nested inserts show complex or even domain-like fold patterns with helices, strands and loops. Furthermore, a basal level of structural innovation was found in inserts which displayed a significant structural similarity exclusively to themselves. The beta-Lactamase/D-ala carboxypeptidase domain family is provided as an example to illustrate the inference of insertion events, and how the incremental growth of a variable region is capable to generate novel structural patterns. CONCLUSION: Using 3D data, we proposed a method to reconstruct phylogenies. We applied the method to reconstruct the sequences of insertion events leading to the emergence of potentially novel structural elements within existing protein domains. The results suggest that structural innovation is possible via the stochastic process of insertions and rapid evolution within variable regions where inserts tend to be nested. We also demonstrate that the structure-based phylogeny enables the study of new questions relating to the evolution of protein domain and biological function.

BACKGROUND: An increasing number of bioinformatics methods are considering the phylogenetic relationships between biological sequences. Implementing new methodologies using the maximum likelihood phylogenetic framework can be a time consuming task. RESULTS: The bioinformatics library libcov is a collection of C++ classes that provides a high and low-level interface to maximum likelihood phylogenetics, sequence analysis and a data structure for structural biological methods. libcov can be used to compute likelihoods, search tree topologies, estimate site rates, cluster sequences, manipulate tree structures and compare phylogenies for a broad selection of applications. CONCLUSION: Using this library, it is possible to rapidly prototype applications that use the sophistication of phylogenetic likelihoods without getting involved in a major software engineering project. libcov is thus a potentially valuable building block to develop in-house methodologies in the field of protein phylogenetics.