Instance maps as an organising concept for complex experimental workflows as demonstrated for (nano)material safety research

Abstract

Nanosafety assessment, which seeks to evaluate the risks from exposure to nanoscale materials, spans materials synthesis and characterisation, exposure science, toxicology, and computational approaches, resulting in complex experimental workflows and diverse data types. Managing the data flows, with a focus on provenance (who generated the data and for what purpose) and quality (how was the data generated, using which protocol with which controls), as part of good research output management, is necessary to maximise the reuse potential and value of the data. Instance maps have been developed and evolved to visualise experimental nanosafety workflows and to bridge the gap between the theoretical principles of FAIR (Findable, Accessible, Interoperable and Re-usable) data and the everyday practice of experimental researchers. Instance maps are most effective when applied at the study design stage to associate the workflow with the nanomaterials, environmental conditions, method descriptions, protocols, biological and computational models to be used, and the data flows arising from study execution. Application of the InstanceMaps tool (described herein) to research workflows of increasing complexity is presented to demonstrate its utility, starting from (i) documentation of a nanomaterial's synthesis, functionalisation, and characterisation, over (ii) assessment of a nanomaterial's transformations in complex media, (iii) description of the culturing of ecotoxicity model organisms Daphnia magna and their use in standardised tests for nanomaterials ecotoxicity assessment, and (iv) visualisation of complex workflows in human immunotoxicity assessment using cell lines and primary cellular models, to (v) the use of the instance map approach for the coordination of materials and data flows in complex multipartner collaborative projects and for the demonstration of case studies. Finally, areas for future development of the instance map approach and the tool are highlighted.

Keywords: FAIR; data collection and quality control; data provenance; experimental workflow visualisation; nanomaterial life cycle stages; study design.

Figure 1

Comparison between (a) the original concept of an instance map using the original definition from NIKC, modified from Amos et al. [27] and (b) an instance map generated using the InstanceMaps tool with its extended node library. The full instance map in (b) is available at https://figshare.com/articles/software/25416040?file=51103502 for interactive inspection. (c) Comparison of the categories of instance map nodes between the original version and the InstanceMaps tool and illustration of the new features available via the InstanceMaps tool.

Figure 2

The nodes available in the InstanceMaps tool to represent a study, grouped into four categories. An instance consists of the material and its medium (surroundings). Properties can be curated (from literature) or calculated (computed) or experimentally determined. Protocols cover all steps of the workflow, including any transformations, sample dispersion and exposure, measurement steps (e.g., physicochemical characterisation, (eco)toxicity evaluation, and functional testing), and data processing such as gap-filling, data cleaning, and statistical analyses. Data is then classified as raw (coming directly from the measurement) or processed (following steps such as subtraction of medium blanks or calculation of half maximal effect concentrations).

Figure 3

This part of an instance map shows the first steps of the synthesis protocol for the sulfidation of AgNPs originally published in [34]. The full instance map is shown as a miniature on the top right and is available at https://figshare.com/articles/software/25416040?file=51103502 for interactive inspection. The map can be divided in two phases. The first phase is the synthesis of silver nanoparticles (AgNPs) functionalised with polyvinylpyrrolidone (PVP), which occurs in the first three instances (shown as light blue nodes). Once the NPs have been synthesised, physical and chemical characterisation of the particles is performed. These characterisation endpoints can be seen in orange linked to the material in the third instance. The second phase is the sulfidation process, which can be seen in the light purple boxes (only accessible in the interactive tool). Although there are three “transformation protocols” listed, the protocol is the same except for the concentration of PVP-AgNPs used.

Figure 4

Instance map of a nanomaterial’s mesocosm experiment. (A) Representation of an instance map for a mesocosm exposure experiment. (B) An expanded map region to visualise the experiment organisation and the flow of data collection. Instances (blue boxes outlined with a blue border) are organised in time from the top of the map (blue arrow represents direction of time). Data is split on either side of the instances to distinguish its origin. To the left are the green and yellow boxes that show curated data and pre-experiment information. Curated data and pre-experiment information is further split across instances to show when it is applicable to pre- (green) or post- (yellow) exposure of organisms. To the right side of the instances is an orange box that shows all data generated from a given instance. This data is also further split into two categories. First, raw data and processed data (pink border) and, second, the methodologies and processing approaches used to derive that data (purple border). (C) Extension of a sample node to include further analysis and data points. The full instance map is available at https://figshare.com/articles/software/25416040?file=51103502 for interactive inspection.

Figure 5

Instance map visualising the steps in maintaining continuous D. magna cultures. Daphnia typically produce brood from about ten days of age and roughly every three days thereafter, with the third to seventh broods being the most genetically stable and, thus, suitable for ecotoxicity experiments. Tracking of the number of offspring per brood is one of the essential QC measures to record, using the template shown in Table 2. Details such as organism species, strain, and culturing conditions (temperature, pH, dissolved oxygen, light/dark cycle) can be captured here as well as the specifics of, for example, the medium and the culturing vessels. The full instance map is available at https://figshare.com/articles/software/25416040?file=51103502 for interactive inspection.

Figure 6

Representation of the OECD 202 “Daphnia sp. acute immobilisation test” guideline for acute toxicity to daphnia as an instance map. For nanomaterials, there would be an additional link from the stock solution to the range of characterisation studies needed, such as size distribution, surface charge, and stability over time. The full instance map is available at https://figshare.com/articles/software/25416040?file=51103502 for interactive inspection.

Figure 7

Representation of the OECD 211 “Daphnia magna reproduction test” guideline for reproductive (chronic) toxicity to daphnia as an instance map (b). The exposure concentration is determined from the acute dose–response curve (a), generated according to the instance map shown in Figure 6, which will be fully integrated in a next iteration of the InstanceMaps tool. The instance map is available at https://figshare.com/articles/software/25416040?file=51103502 for interactive inspection.

Figure 8

Instance map of the immunotoxicity workflow to study the bio-nano interactions of differently functionalised SiO₂ nanomaterials with immune cells. The instance map is divided into sections A–E, based on the studies of Hasenkopf et al. [48], Mills-Goodlet et al. [49], Johnson et al. [50], and Punz et al. [51], highlighting the different approaches and routes that were taken. Section A, which serves as a baseline for all studies, mainly focuses on the nanomaterial synthesis and surface modification. The pathway towards more in-depth immunological investigations was chosen for section B, while sections C and E cover alterations in the protein binding activity, depending on the physicochemical properties provided by chemical surface functionalisation and also the observed structural alterations that occurred upon nanomaterial conjugation because of the nanotopography of the materials (mesoporous SiO₂ nanomaterials). Section D depicts the integration of in silico predictive modelling approaches with quantitative and qualitative in vitro determination of the protein corona (epitope rearrangement). The full instance map is available at https://figshare.com/articles/software/25416040?file=51103502 for interactive inspection.

Figure 9

Part of the instance map depicting the planning status of the human and ecotoxicology testing for the MACRAMÉ use case of antibiotics-loaded polymeric nanomaterials. After the production of the loaded nanomaterials, they are sent to many experimental partners performing the different assays. The instance map is crucial to describe the complexity of the workflow, which includes strong cross-partner dependencies such as sample preparation by one partner and measurement by another, which must be completed within a specific timeframe. The full instance map is available at https://figshare.com/articles/software/25416040?file=51103502 for inspection.

Figure 10

Part of the instance map for the MACRAMÉ use case of antibiotic-loaded nanoparticles representing the shipping of the pristine material to partners performing the human and ecotoxicity testing. The full instance map is available at https://figshare.com/articles/software/25416040?file=51103502 for interactive inspection.