SIGLASEARCH: arcabouço de consulta de dados flexibilizada para workflows de múltiplos laboratórios de pesquisa

Abstract

Biomedical research has undergone strong growth of data generated by the expansion of the law governing laboratory procedures, advances in automated instrumentation, improved quality control processes and increased demand for statistical analysis. To control this volume of information, some research labs use Laboratory Information Management Systems (LIMS), which propose recording management and access to data safely and satisfactorily. Research labs have quite peculiar features and work processes. Therefore, the LIMS are created with custom data structure for each of them. This need for software adaptation increases development time and cost of the final application, which makes the implementation of a LIMS impractical for small laboratories. In order to increase availability of LIMS for the laboratory, the Integrated Laboratory Management System (Sistema Integrado de Gerência de Laboratórios - SIGLa) was created, which enables the laboratory to adapt the data flow of the system according to their business needs without changes in the source code of the system. To change the work processes of a lab, just the workflow is edited through a workflow editor such as the Together Workflow Editor (TWE). SIGLa acts as a workflow manager, making the transaction control, data consistency and data recording. SIGLa is already used in the Brazilian Agricultural Research Corporation (Empresa Brasileira de Pesquisa Agropecuária - Embrapa) and the National Institute of Metrology, Quality and Technology (Instituto Nacional de Metrologia, Qualidade e Tecnologia - Inmetro). Although data recording is flexible and does not require customizations for change of workflows, the query data is not satisfactory. The existing query displays the recorded data, but displays no such related form of data. This problem has a nontrivial solution since the system modeling has been made in Relational Database according to the concepts of workflow manager, such as activities, attributes, records and instances. This implies low value of the results of the old query system for information analysis. The tables are related by their metadata, so the user can not extract the relationships between the activities of the workflow, relationships between the attributes of these activities and relationships between records that make some types of workflow attributes and other attributes. Tools existing today for automatic query Relational Databases, do not solve the query problem in SIGLa, because the extraction of the constraints of relational schema is not enough to generate the results that the user needs in his analysis. This project proposes the modeling and implementation of SiglaSearch, a system to query workflow data with customized user interface to different users, more flexible for different workflows, to view the information related with each other without the need for conversion of data model-driven workflow to a model generated in terms of the relationships between the data. The new query system developed for SIGLa brings as improvement research in workflow data, because the existing reference systems for such data only offer keyword search while SiglaSearch proposes to search through a subset of workflow information. SiglaSearch allows the search of data subsets, filtering and sorting of any workflow information and summarization of results. Relationships are extracted automatically without additional specifications, with display of information in tabular format, so users can focus on subsets of information to facilitate the evaluation of information for various purposes. SiglaSearch also provides a query interface for users so that they can get their results in a few steps intuitively and can save these consultations in the form of reports that can be run by other users working in more specific activities of the laboratory search.