Webinar: From Project to Platform: New Resources on PSDI – Session 6

Challenge

• • AFM generates rich multimodal nanoscale data without the need for labelling or vacuum environments.
  • Quantitative analysis has historically been limited by manual, non-standardised, and difficult-to-reproduce workflows.
  • Limited availability of automated analytical tools and machine learning integration has restricted large-scale, data-driven AFM research.

Approach

• • AFM Reader
    • FAIR open-source utility for standardised access to AFM datasets across multiple instrument formats.
    • Simplifies data loading and improves interoperability between datasets.
  • TopoStats
    • FAIR open-source Python pipeline for data cleaning, processing, feature identification, and instance segmentation.
    • Supports automated batch analysis, enabling robust and repeatable quantitative measurements.
  • AFMDB
    • Open repository linking raw AFM files (.spm, .jpk, .asd, .ibw, etc.) with processed outputs.
    • Promotes open data practices and long-term reuse of AFM datasets.

Innovation

• • Enables scalable, fully quantitative, and reproducible AFM data analysis through an integrated open-source ecosystem.
  • Combines data access, automated image analysis, and data management within a FAIR framework.
  • Creates a growing repository of curated AFM datasets that:
    • facilitates data sharing and community benchmarking;
    • supports the training and validation of machine learning and deep learning models;
    • accelerates the development of next-generation AI-enabled AFM analysis tools.