NFDI4BIOIMAGE

NFDI4BIOIMAGE Carsten Fortmann-Grote1 April 18 2024 This presentation is released to the public domain under Creative Commons Attribution License CC-BY-4.0 International. 1ca[email protected] Carsten Fortmann-Grote NFDI4BIOIMAGE April 18 2024 1 / 24 The NFDI4BIOIMAGE Consortium •Harmonization of bioimage data and metadata formats and standards •FAIR Image Objects •Multimodal metadata vocabularies •Standard Operating Procedures for FAIR image processing •Proliferation of best practices in image data management and processing Carsten Fortmann-Grote NFDI4BIOIMAGE April 18 2024 2 / 24 All Hands Meeting Oct. 2023 Carsten Fortmann-Grote NFDI4BIOIMAGE April 18 2024 3 / 24 No Milestones due yet Carsten Fortmann-Grote NFDI4BIOIMAGE April 18 2024 4 / 24 Outline •FAIRification of image data: 5 star scheme •Teaching & Training material •BIDS, ARC, OMERO interfaces •Data integration at MPI-EvolBio Carsten Fortmann-Grote NFDI4BIOIMAGE April 18 2024 5 / 24 FAIRification of Image (meta)data Quantify the FAIRness of containerized Bioimage research objects Carsten Fortmann-Grote NFDI4BIOIMAGE April 18 2024 6 / 24 FAIRification of Image (meta)data 5 stars linked open data Carsten Fortmann-Grote NFDI4BIOIMAGE April 18 2024 7 / 24 FAIRification of Image (meta)data 5 star bioimage containers Carsten Fortmann-Grote NFDI4BIOIMAGE April 18 2024 8 / 24 Zarr is the technical backbone of next generation (bio)image file formats What is Zarr? •hdf5 like hierarchical layers •chunks stored in separate "files" (blobs), linked by json file (vs. monolithic binary blob in hdf5) •suitable for object storage dataset.tif [120 GB] .zarray [240 B] 0.0.0 [260 kB] 0.0.1 [260 kB] 0.0.2 [260 kB] 10010100100001001011 10101001010010100101 00100101101001001111 01110101010100010010 00000010100010010100 01001001011101001001 00101110111111101011 1010101010101000010010101010101010000100 10010110010100100001 00101110101001010010 10010100100101101001 00111101110101010100 01001000000010100010 01010001001001011101 ACCESS ACCESS ACCESS { "dtype": " | u1", . . . } Monolithic file format OME-Zarr 3000 60 um 100 um 90 um X Y Z 10.5281/zenodo.7037679 Carsten Fortmann-Grote NFDI4BIOIMAGE April 18 2024 9 / 24 Many open questions •One-to-one mapping of ARC entities to/from OMERO? •What about non-image data in ARC? •OMERO as metadata manager on top of ARC? •round trip possible? •large image files (think TB) in git? Carsten Fortmann-Grote NFDI4BIOIMAGE April 18 2024 15 / 24 Mapping (attempts) DOI:10.5281/zenodo.8349563 Carsten Fortmann-Grote NFDI4BIOIMAGE April 18 2024 16 / 24 Conclusions so far •conversion OMERO →BIDS →ARC seems possible •only for already specified measurements (BIDS standard) •requires constant reformatting (e.g. fastq to csv) •⇔Support by format specific processing clients Carsten Fortmann-Grote NFDI4BIOIMAGE April 18 2024 17 / 24 MPI-EvolBio’s activities Evolutionary Genetics (D. Tautz (Emeritus)) •Model organisms: Mus domesticus, mus musculus •Behavioural genomics •Population genetics credits: https://www.evolbio.mpg.de/ 3039130/group_evolanimalbehpers Microbial Population Biology (P. Rainey) •Model organism: Pseudomonas fluorescens, Bacillus sub. •Evolution of communities •Host-microbe interactions •Genetics credits: Theodosiou (left), Schwarz (middle), Grote (right) Theoretical Biology (A. Traulsen) •Population Structure and Game Theory •Metaorganisms •Cancer Evolution Hindersin et al, PLOS Comp. Bio (2019) 10.1371/journal.pcbi.1004437 Carsten Fortmann-Grote NFDI4BIOIMAGE April 18 2024 18 / 24 Data Management Policy enforces deposition of all research data Data collection Data analysis DMP Sequence data Materials Notes Protocols Simulation data Image data Code OMERO Storage openBIS GitLab/Hub DMP FDK DMP Library FDK edmondzenodo ENABIA GitLab/Hub PuRe Results Publication Library Reuse Checklist Carsten Fortmann-Grote NFDI4BIOIMAGE April 18 2024 19 / 24 Multimodal, multidimensional data in experimental evolution research •Wildtype clone and/or genetically modified strain(s) •Timelapse microscopy from growing microbial communities •Time resolved whole genome / core gene NGS data •Transcription profiles •Optical density measurements from plaque assays •Functional annotations Need for integrated analysis of multiomics-multimodal data "Find images, ELN entries for ∆mreB strains and annotations for mreB" Carsten Fortmann-Grote NFDI4BIOIMAGE April 18 2024 19 / 24 Integrating data the SPARQLing way Carsten Fortmann-Grote NFDI4BIOIMAGE April 18 2024 20 / 24 Data integration of internal sources DB API LOD ready? RDF conversion Tripal Genome DB JSON-LD yes SPARQL SPARQL-anything virtualization JSON to RDF serialization OMERO JSON-LD yes omero-rdf OpenBIS JSON-LD yes ??? ("semantic annotation") StrainDB None no csv dump →csv2rdf Carsten Fortmann-Grote NFDI4BIOIMAGE April 18 2024 21 / 24 Flattening a deeply nested JSON-LD graph •rdflib for loading json-ld into graph structure •iteratively parses linked graph uris •dask for distributed computing •serialized to turtle format •-> 2.1 million Triples in ca. 12hrs Carsten Fortmann-Grote NFDI4BIOIMAGE April 18 2024 22 / 24 SparNatural for visual query generation Carsten Fortmann-Grote NFDI4BIOIMAGE April 18 2024 23 / 24