BackgroundPersonalized medicine has become a priority in breast cancer patient management. In addition to the routinely used clinicopathological characteristics, clinicians will have to face an increasing amount of data derived from tumor molecular profiling. The aims of this study were to develop a new gene selection method based on a fuzzy logic selection and classification algorithm, and to validate the gene signatures obtained on breast cancer patient cohorts.MethodsWe analyzed data from four published gene expression datasets for breast carcinomas. We identified the best discriminating genes by comparing molecular expression profiles between histologic grade 1 and 3 tumors for each of the training datasets. The most pertinent probes were selected and used to define fuzzy molecular grade 1-like (good prognosis) and fuzzy molecular grade 3-like (poor prognosis) profiles. To evaluate the prognostic performance of the fuzzy grade signatures in breast cancer tumors, a Kaplan-Meier analysis was conducted to compare the relapse-free survival deduced from histologic grade and fuzzy molecular grade classification.ResultsWe applied the fuzzy logic selection on breast cancer databases and obtained four new gene signatures. Analysis in the training public sets showed good performance of these gene signatures for grade (sensitivity from 90% to 95%, specificity 67% to 93%). To validate these gene signatures, we designed probes on custom microarrays and tested them on 150 invasive breast carcinomas. Good performance was obtained with an error rate of less than 10%. For one gene signature, among 74 histologic grade 3 and 18 grade 1 tumors, 88 cases (96%) were correctly assigned. Interestingly histologic grade 2 tumors (n = 58) were split in these two molecular grade categories.ConclusionWe confirmed the use of fuzzy logic selection as a new tool to identify gene signatures with good reliability and increased classification power. This method based on artificial intelligence algorithms was successfully applied to breast cancers molecular grade classification allowing histologic grade 2 classification into grade 1 and grade 2 like to improve patients prognosis. It opens the way to further development for identification of new biomarker combinations in other applications such as prediction of treatment response.Electronic supplementary materialThe online version of this article (doi:10.1186/s12920-015-0077-1) contains supplementary material, which is available to authorized users.

Autor a quien debe ser dirigida la correspondencia Recibido Sep. 12, 2012; Aceptado Nov. 06, 2012; Versión final recibida Dic. 15, 2012 Resumen Este artículo presenta una metodología para predecir estados funcionales en procesos complejos a partir de la estimación de grados de pertenencia difusos. La propuesta integra una medida estática como es el resultado de un clasificador difuso entrenado con los datos históricos del proceso y un algoritmo de estimación basado en la teoría de Markov para eventos discretos. La propuesta, que puede ser integrada a un sistema de monitoreo de sistemas complejos, comprende dos etapas: una etapa de entrenamiento fuera de línea para definir el clasificador difuso y el estimador; y una etapa en línea donde se realizan la clasificación de la situación actual del proceso y la estimación del estado funcional para el siguiente tiempo de muestreo. La propuesta desarrollada para la estimación de estados funcionales permite utilizar cualquier método de agrupamiento difuso que suministre la información base que requiere la metodología. La metodología fue probada con éxito en un sistema de monitoreo para una línea de transmisión de energía y en el monitoreo de un sistema de caldera. Palabras clave: predicción de estados funcionales, clasificador difuso, agrupamiento difuso, cadenas de Markov AbstractThis paper presents a methodology to predict functional states in complex processes from the estimation of fuzzy membership degrees. The proposal integrates a static measure, such as the result of a fuzzy classifier trained with historical process data, and an estimation algorithm based on Markov theory for discrete events. The proposal, which can be integrated to the monitoring of complex systems, provides two stages: an off-line training stage to define the fuzzy classifier and the estimator; and an online stage where the classification of the current process situation and the estimation of the next functional state are performed. The proposal for the estimation of functional states allows using any fuzzy clustering method that provides the information required by the methodology. The proposed methodology was successfully tested on a monitoring system for a power transmission line and in the monitoring of a boiler system.