Implementing interoperable provenance in biomedical research

Abstract

The provenance of a piece of data refers to knowledge about its origin, in terms of entities and actors involved in its creation, e.g. data sources used, operations carried out on them, and users enacting those operations. Provenance is used to better understand the data and the context of its production, and to assess its reliability, by asserting whether correct procedures were followed. Providing evidence for validating research is of particular importance in the biomedical domain, where the strength of the results depends on the data sources and processes used. In recent times, previously manual processes have become fully or semi-automated, e.g. clinical trial recruitment, epidemiological studies, diagnosis making. The latter is typically achieved through interactions of heterogeneous software systems in multiple settings (hospitals, clinics, academic and industrial research organisations). Provenance traces of these software need to be integrated in a consistent and meaningful manner, but since these software systems rarely share a common platform, the provenance interoperability between them has to be achieved on the level of conceptual models. It is a non-trivial matter to determine where to start in making a biomedical software system provenance-aware. In this paper, we specify recommendations to developers on how to approach provenance modelling, capture, security, storage and querying, based on our expe- * riences with two large-scale biomedical research projects: Translational Research and Patient Safety in Europe (TRANSFoRm) and Electronic Health Records for Clinical Research (EHR4CR). While illustrated with concrete issues encountered, the recommendations are sufficiently high level so as to be reusable across the biomedical domain.

Static analysis of Taverna workflows to predict provenance patterns

“…Similarly given a task with an iteration strategy in the workflow, each invocation of task will be documented distinctly in provenance. We shall note that recent research on provenance have introduced the notion of "provenance patterns/templates" [11] [12]. Templates denote fragments of provenance that are expected to be collected by a system, and they are commonly represented in standard models such as OPM or PROV.…”

“…Starting with desired provenance patterns and deducing the workflow that would generate those patterns. This would require the specification of anticipated provenance patterns, a thread of research which is currently being explored with two research groups [12] [11]. Here a standard provenance language such as PROV is being enriched with variables to represent patterns abstracted from any particular data or process identity.…”

“…In case such patterns can be represented as abstract provenance graphs then (workflow) system specific techniques can be devised in order to map patterns to workflows that can generate such patterns. We shall note that the current specifications for abstract provenance [11] [12] are still under development as this research is very recent. We conjecture that abstract provenance representations would potentially require extending with models to represent data collections and constraint languages to represent patterns, such as traceability among collections, or patterns that depend on data values.…”

Developing provenance-aware query systems: an occurrence-centric approach

“…The work of Glavic and Dittrich [21] categorizes provenance systems by the functionalities (established as a set) that it supports. In a similar vein, Curcin et al [14] present a list of twenty-one recommendations outlining the ideal provenance systems. These recommendations are grouped into five categories, namely modeling, capture, storage, security, and querying and visualization.…”

Drug discovery FAQs: workflows for answering multidomain drug discovery questions

“…Most scientific workflow programs, such as KNIME [7], Taverna [9], or Pipeline Pilot [10], provide a userfriendly graphical workbench that enables scientists to create and visualize complex workflows easily that might comprise dozens of processing and analytical steps. Furthermore, many workflows provide mechanisms for tracing provenance and other methodologies that foster reproducible science [11,12]. Typically, the functionality of the workflow components falls into various categories, such as data transformation, data preparation and data analysis [13], and can be both part of the workbench or provided via web services (as application programming interfaces or APIs).…”