Resource Themes

The PSDI Knowledge Base offers guidance and training materials for physical scientists covering various themes related to working with scientific research data, and using the technologies and resources provided by PSDI. The main collections of topics provided are described in this resource theme.

BioSimDR (Biomolecular Simulation Data Resources) is a collection of tools and data services designed to transform and store biomolecular simulations by capturing and archiving the full provenance of each simulation. Our tools can be used to meticulously record every step - from experimental structures to force field configurations and computational parameters - ensuring research is reproducible, transparent, and efficient. Our data services allow for biomolecular simulations to be stored, shared, and discovered in one place. This approach helps scientists to avoid redundant efforts, improves the reliability of findings, and supports AI-driven drug discovery and molecular modelling.

The CSD is the world's largest curated database of real-world, small molecule organic and metal-organic crystal structures. All data are reviewed and annotated by expert scientists to enable easy retrieval and subsequent reuse in a range of contexts. The CSD can be used to guide molecular and materials design by comparing features of new molecules and materials against what has been experimentally observed to identify areas of risk or strategies for optimisation. The CSD can be searched via the PSDI Cross Data Search service allowing results to be combined with other data sources to generate novel insights to guide the design of new molecules and materials.

This resource presents data exemplars and detailed case study reports from our investigations into unlocking the potential of legacy data. It highlights our approaches to meaningfully revive outdated datasets, enhancing their quality, and maximising their value. These efforts aim to provide practical guidance on methods for resurrecting data, uncovering value in hidden data sources, and accelerating scientific progress.

The Collaborative Computational Project for NMR Crystallography (CCP-NC) supports a multidisciplinary community by developing and integrating software tools that advance NMR crystallography. By combining experimental NMR with computational methods, CCP-NC enables researchers to gain atomic-scale insights into structure, disorder, and dynamics in the solid state. A key resource is its repository for first-principles NMR calculations, which standardises data using the .magres file format. The project also provides specialised software, such as MagresView, for data processing and visualisation, along with tools for electronic structure calculations and ab initio NMR parameter prediction. Additionally, CCP-NC offers training opportunities and access to high-performance computing (HPC) through the Young supercomputer, further enhancing research capabilities. Through these resources, CCP-NC empowers researchers to conduct cutting-edge NMR crystallography, improving efficiency, fostering collaboration, and expanding professional networks. By facilitating the development and adoption of advanced computational methods, the project contributes to the broader scientific community, driving innovation in materials research.

This resource theme offers access to carefully curated data collections from diverse sources, as well as data exemplars developed through case studies. These exemplars explore innovative methods to enhance data capture, improve quality, facilitate sharing, and revive legacy data in meaningful ways. Our goal is to create, store, and manage robust data collections in the physical sciences, ensuring they are easily accessible to researchers. By simplifying data discovery and reuse, we promote broader utilisation and accelerate scientific advancements.

Machine learning has transformed the field of materials modelling in the last few years. Given access to high-quality data from computations and/or experiments, machine learning can be used to develop expert systems powered by large language models (such as ChatGPT). These expert systems can be to train surrogate models that can predict properties of structures, eliminating the need for simulations, or speed up simulations by using machine learnt interatomic potentials (MLIPs). The Data to Knowledge resource theme is dedicated to making the creation of these models possible by providing data infrastructure and workflows enabling the generation and exploitation of these machine learnt models. The Data to Knowledge collections comprise curated datasets designed for use in machine learning or generated through machine learning. An example is the Machine Learning Interatomic Potentials (MLIPs) data collection, which includes MLIPs XYZ files used for training, the trained model itself, and, where possible, related data such as AIIDA provenance records. Making these datasets available enables those without the resources to compute data themselves to utilise them for machine learning and modelling. Training is a central focus of this resource theme. We provide two types of training: general training and tool-specific training. Our general training is provided as self-paced learning online.

Post-processing experimental and simulation data from large-scale experiments conducted at national facilities requires integrating diverse software tools from various domains into cohesive workflows. Although these workflows can be highly complex, the software tools and data associated with them can be effectively managed using the open-source, web-based Galaxy Workflow Management System. Designed for FAIR data analysis, Galaxy enables users to run code in interactive environments, share and publish results, workflows, and visualisations, and ensure reproducibility by capturing and packaging the data, metadata, and provenance models essential for replicating and understanding their analyses. Use cases are provided for muon science and catalysis experiments, with plans to expand to other large-scale experiments.

Your one-stop shop for file format conversion, PSDI's Data Conversion Service enables users to seamlessly convert between different file formats and assess the quality of proposed conversions. Several options are available, depending on whether you want to upload your data to PSDI or run the conversion locally or as part of your own applications. These options include: the PSDI Hosted Web Application, a standalone Command Line application that can be installed and run locally, and a Python library for integration into your own pipelines.

The majority of laboratories and research groups have stacks of paper notebooks filled with valuable data that is difficult and time-consuming to access and use. Data Revival is an AI-driven service that can seamlessly convert your handwritten lab book pages into machine-readable data, unlocking the potential of your chemistry data. Operating with a high degree of accuracy, Data Revival creates relational links across your chemical data and enables users to run semantic searches across potentially millions of data points, including complex chemical structures. PSDI offers users the chance to experience Data Revival's capabilities firsthand.

This resource theme provides a range of services and guidance to help users explore and appreciate the importance of digital notebooks in documenting their experiments and research. The guidance introduces the benefits of both generic digital notebooks and Electronic Laboratory Notebooks (ELNs), and the different considerations for implementation in laboratories. The resource theme includes case studies and exemplars. It also links to ELNFinder which is an external service for which PSDI is an active collaborator. This service helps users find the most suitable ELN based on a comprehensive set of criteria.

Today's researchers must possess skills that extend beyond the specialised knowledge gained through their degrees and professional roles. The Skills4Scientists guidance materials are designed to enhance researchers' capabilities in areas such as computing, research data management, and other transferable research skills, supporting them in all aspects of their work. These resources include interactive learning materials, practical guidance topics, and best practices. Additionally, we are collaborating with the community to continuously grow and improve the guidance offered through PSDI.

The Physical Sciences Data Infrastructure (PSDI) is set to become the largest UK data infrastructure for the physical sciences, supporting diverse research communities in accessing and sharing data resources. Community portals are designed to provide customized access points, enabling communities to showcase the research outputs generated using PSDI's resources and services. The database, services, code, and deployed portal demonstrate the benefits of community portals, highlighting their adaptability and ease of customization and deployment for diverse research communities.

Computing environments, including software applications and dependencies, can be packaged into containers to enhance portability, installation, and usability. This resource is a registry of PSDI-developed containers designed for researchers in the physical sciences.

This resource theme includes data sources searchable via the PSDI Cross Data Search service. Users can search for substances (e.g., molecules or crystal structures) or publication references, depending on the selected data sources. It incorporates data from PSDI, partners, OPTIMADE providers, and other relevant external sources, offering a comprehensive discovery tool for the research community.

The Open Databases Integration for Materials Design (OPTIMADE) consortium aims to make materials databases interoperable by developing a specification for a common REST API. This resource theme includes data providers who share their datasets through an OPTIMADE endpoint, making them accessible and searchable via the PSDI Cross Data Search Service.