Análisis visual de datos biofarmacéuticos
Fecha
2017-06-25
Autores
Rodríguez Mollineda, David
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Central “Marta Abreu” de Las Villas
Resumen
La sobrecarga de información causada por el exceso de datos biofarmacéuticos plantea un
reto para el análisis de propiedades relacionadas con el proceso de absorción oral de
fármacos y la determinación de su biodisponibilidad, cuyas características hacen que sea
poco apropiado el uso de técnicas completamente automatizadas y cuyo volumen impide a
los especialistas examinarlos directamente. Este trabajo permitió identificar relaciones y
patrones interesantes dentro de las variables vinculadas a este proceso, empleando métodos
y herramientas del análisis visual de datos como alternativa a otros enfoques más
tradicionales. Para ello, se concibió un modelo de análisis basado en los principios del
análisis visual y ajustado a las características propias de los datos biofarmacéuticos,
centrándose en el empleo de herramientas analíticas como el agrupamiento y técnicas de
visualización destinadas al trabajo con datos multivariados como formas de análisis e
interacción fundamentales. Como validación del modelo se realizó la implementación de
una herramienta informática basada en este, con énfasis en la extensibilidad de sus
funcionalidades y la flexibilidad en su uso. La aplicación de esta a un caso de estudio
relacionado con la extracción de información referente a una de las propiedades objetivo de
optimización en el proceso de desarrollo de fármacos, permitió confirmar la efectividad del
modelo desarrollado y del análisis visual de datos en general, para dar solución a este
problema de análisis.
The information overload caused by the excess of biopharmaceutical data poses a challenge for the analysis of properties related to oral absorption process of drugs and the assessment of their bioavailability, whose characteristics make it inappropriate to apply fully automated techniques as well as their volume impede the specialists from directly examining the data. This work allowed to identify interesting relations and patterns within the variables related to this process, employing Visual Analytics methods and tools as an alternative to more traditional approaches. In order to achieve this, an analysis model was conceived based on the principles of Visual Analytics and adjusted to the characteristics of biopharmaceutical data, focusing in the use of analytical tools such as clustering and visualization techniques meant for working with multivariate data as its main analysis and interaction forms respectively. As validation for the model, an informatics tool based on it was implemented, paying special attention to the extensibility of its functions and the flexibility of its use. The application of said tool to a case of study related to extracting information related to one of the optimization target properties in the drug development process, allowed to confirm the effectiveness of the developed model and Visual Analytics in general, to provide a solution to this analysis problem.
The information overload caused by the excess of biopharmaceutical data poses a challenge for the analysis of properties related to oral absorption process of drugs and the assessment of their bioavailability, whose characteristics make it inappropriate to apply fully automated techniques as well as their volume impede the specialists from directly examining the data. This work allowed to identify interesting relations and patterns within the variables related to this process, employing Visual Analytics methods and tools as an alternative to more traditional approaches. In order to achieve this, an analysis model was conceived based on the principles of Visual Analytics and adjusted to the characteristics of biopharmaceutical data, focusing in the use of analytical tools such as clustering and visualization techniques meant for working with multivariate data as its main analysis and interaction forms respectively. As validation for the model, an informatics tool based on it was implemented, paying special attention to the extensibility of its functions and the flexibility of its use. The application of said tool to a case of study related to extracting information related to one of the optimization target properties in the drug development process, allowed to confirm the effectiveness of the developed model and Visual Analytics in general, to provide a solution to this analysis problem.
Descripción
Palabras clave
Análisis Visual, Biodisponibilidad, Clasificación BCS, Agrupamiento, Sobrecarga de Información, Datos Multivariados, Visual Analytics, Bioavailability, BCS Classification, Clustering, Information Overload