Open Science is key to future scientific research and promotes a deep transformation in the whole scientific research process encouraging the adoption of transparent and collaborative scientific approaches aimed at knowledge sharing. Open Science is increasingly gaining attention in the current and future research agenda worldwide. To effectively address Open Science goals, besides Open Access to results and data, it is also paramount to provide tools or environments to support the whole research process, in particular the design, execution and sharing of transparent and reproducible experiments, including data provenance (or lineage) tracking. This work introduces the Climate Analytics-Hub, a new component on top of the Earth System Grid Federation (ESGF), which joins big data approaches and parallel computing paradigms to provide an Open Science environment for reproducible multi-model climate change data analytics experiments at scale. An operational implementation has been set up at the SuperComputing Centre of the Euro-Mediterranean Center on Climate Change, with the main goal of becoming a reference Open Science hub in the climate community regarding the multi-model analysis based on the Coupled Model Intercomparison Project (CMIP).
Whenever high-performance computing applications meet data-intensive scalable systems, an attractive approach is the use of Apache Spark for the management of scientific workflows. Spark provides several advantages such as being widely supported and granting efficient in-memory data management for large-scale applications. However, Spark still lacks support for data tracking and workflow provenance. Additionally, Spark’s memory management requires accessing all data movements between the workflow activities. Therefore, the running of legacy programs on Spark is interpreted as a "black-box" activity, which prevents the capture and analysis of implicit data movements. Here, we present SAMbA, an Apache Spark extension for the gathering of prospective and retrospective provenance and domain data within distributed scientific workflows. Our approach relies on enveloping both RDD structure and data contents at runtime so that (i) RDD-enclosure consumed and produced data are captured and registered by SAMbA in a structured way, and (ii) provenance data can be queried during and after the execution of scientific workflows. By following the W3C PROV representation, we model the roles of RDD regarding prospective and retrospective provenance data. Our solution provides mechanisms for the capture and storage of provenance data without jeopardizing Spark’s performance. The provenance retrieval capabilities of our proposal are evaluated in a practical case study, in which data analytics are provided by several SAMbA parameterizations.
There has been a recent major upsurge in the concerns about reproducibility in many areas of science. Within the neuroimaging domain, one approach is to promote reproducibility is to target the re-executability of the publication. The information supporting such re-executability can enable the detailed examination of how an initial finding generalizes across changes in the processing approach, and sampled population, in a controlled scientific fashion. ReproNim: A Center for Reproducible Neuroimaging Computation is a recently funded initiative that seeks to facilitate the ‘last mile’ implementations of core re-executability tools in order to reduce the accessibility barrier and increase adoption of standards and best practices at the neuroimaging research laboratory level. In this report, we summarize the overall approach and tools we have developed in this domain.
Multiple mazes are routinely used to test the performance of animals because each has disadvantages inherent to its shape. However, the maze shape cannot be flexibly and rapidly reproduced in a repeatable and scalable way in a single environment. Here, to overcome the lack of flexibility, scalability, reproducibility and repeatability, we develop a reconfigurable maze system that consists of interlocking runways and an array of accompanying parts. It allows experimenters to rapidly and flexibly configure a variety of maze structures along the grid pattern in a repeatable and scalable manner. Spatial navigational behavior and hippocampal place coding were not impaired by the interlocking mechanism. As a proof-of-principle demonstration, we demonstrate that the maze morphing induces location remapping of the spatial receptive field. The reconfigurable maze thus provides flexibility, scalability, repeatability, and reproducibility, therefore facilitating consistent investigation into the neuronal substrates for learning and memory and allowing screening for behavioral phenotypes.
Recently, many workflows and tools that aim to increase the reproducibility and replicability of research findings have been suggested. In this review, we discuss the opportunities that these efforts offer for the field of developmental cognitive neuroscience. We focus on issues broadly related to statistical power and to flexibility and transparency in data analyses. Critical considerations relating to statistical power include challenges in recruitment and testing of young populations, how to increase the value of studies with small samples, and the opportunities and challenges related to working with large-scale datasets. Developmental studies also involve challenges such as choices about age groupings, modelling across the lifespan, the analyses of longitudinal changes, and neuroimaging data that can be processed and analyzed in a multitude of ways. Flexibility in data acquisition, analyses and description may thereby greatly impact results. We discuss methods for improving transparency in developmental cognitive neuroscience, and how preregistration of studies can improve methodological rigor in the field. While outlining challenges and issues that may arise before, during, and after data collection, solutions and resources are highlighted aiding to overcome some of these. Since the number of useful tools and techniques is ever-growing, we highlight the fact that many practices can be implemented stepwise.
Increasingly complex statistical models are being used for the analysis of biological data. Recent commentary has focused on the ability to compute the same outcome for a given dataset (reproducibility). We argue that a reproducible statistical analysis is not necessarily valid because of unique patterns of nonindependence in every biological dataset. We advocate that analyses should be evaluated with known-truth simulations that capture biological reality, a process we call “analysis validation.” We review the process of validation and suggest criteria that a validation project should meet. We find that different fields of science have historically failed to meet all criteria, and we suggest ways to implement meaningful validation in training and practice.