Poster 28: LiSIs: An Online Scientific Workflow System for Life Sciences ResearchC. C. Kannas1, K. G. Achilleos1, Z. Antoniou1, C. A. Nicolaou1, C. S. Pattichis1, I. Kalvari2, I. Kirmitzoglou2, V. J. Promponas2
|1Department of Computer Science, University of Cyprus|
2Department of Biological Sciences, University of Cyprus
|Scientific Workflow Management Systems (SWMSs) are powerful tools with far reaching potential to facilitate the design and execution of computational experiments. The main objective of SWMSs is to enable scientists to plug together problem solving computational components to implement complex in silico experiments and algorithms requiring multiple computationally intensive steps. Consequently, these tools can accelerate scientific discovery by incorporating data management, analysis, simulation, and visualization tools into a common platform and, by providing the mechanism for pipelining various components to form a workflow custom for a target problem. Typically, SWMSs are equipped with an interactive visual interface that facilitates the design and execution of workflows.|
In recent years, scientific workflow technology has gained popularity in various scientific fields including life sciences research. Numerous SWMSs have paid particular attention to the needs of pharmaceutical researchers and several of them, both commercial and open source, provide access to large collections of computational components and enjoy a size-able user community.
In this presentation we introduce the LIfe Science InformaticS (LiSIs) system, a new, open SWMS, with some unique features designed to enhance user experience and facilitate user adoption. Contrary to most other SWMSs which are desktop-based with back-end server support when needed, LiSIs is an online system based on the widely popular Galaxy SWMS. As such, users access and interact with the system using a regular browser and need not install any software component on their machines. Workflow execution is handled by a dedicated server which may distribute the computational workload appropriately. Moreover,system upgrades are handled centrally when new releases become available in a transparent to the user manner. Team work is supported through the support of online group designation and mechanisms for immediate data and workflow sharing. Provenance is recorded through ‘histories’ associated with specific workflow execution and thereby enable tracing of computational experiments executed and facilitate understanding among team members. LiSIs currently offers numerous computational components for chemical structure property generation, compound fingerprint calculation, predictive modeling, conformation generation and docking among others.
Following the description of the system and its individual components we present results from an application in chemoprevention research. Specifically, LiSIs has been used to implement a virtual screening workflow for the selection of compounds that may serve as leads for subsequent cancer chemoprevention research. Several thousand commercially available compounds were supplied as input to the workflow and were subjected to a series of computational filters including drug likeness, predicted potency via predictive models and predicted binding affinity via docking. The results, shared with expert chemopreventive researchers using the LiSIs platform, demonstrate the potential use of the system by users of varying backgrounds and computational experience to advance drug discovery research.