LAUSR

To the Editor — The cloning and subsequent engineering of fluorescent proteins (FPs) has revolutionized methods in biological research1,2. There are now thousands of FPs available, each with a different combination of properties that affect its usefulness as a genetic tag in various organelles, cell types, model organisms, and experiments. The choice of FP can be extremely consequential for the outcome of an experiment3,4, but currently researchers must pore over the large and ever-expanding body of primary literature to make an informed decision. The magnitude of that task incentivizes the reuse of pre-existing, sometimes suboptimal, FPs. A centralized database of FP characteristics would facilitate the dissemination and adoption of new FPs, and encourage the selection of FPs whose strengths are well matched to the researcher’s application. Large-scale aggregation of FP data might additionally facilitate discovery and inferences about FP properties. FPbase (https://www.fpbase.org) is a free, open-source, web-based, communityeditable database for FPs and their properties. The primary objective is to aggregate structured and searchable FP data that are of interest to the imaging community and FP developers. Each protein in the database has a dedicated page (Fig. 1) showing the amino acid sequence, relevant accession codes (e.g., GenBank, UniProt), evolution lineages and mutations, fluorescence attributes, structural data, references that introduced or characterized the protein, and more. Each protein can have multiple ‘states’ and ‘state transitions’, which allows for the storage of photochromic or environmentally sensitive proteins. Excerpts from primary literature can be entered to store key information about a protein that is otherwise difficult to capture within the current database schema (Supplementary Note 1). In addition to cataloguing FP properties, FPbase includes tools designed to help researchers find and optimize their use of FPs. The advanced search feature filters FPs on the basis of more than 20 characteristics (Supplementary Fig. 1), a sortable table enables users to browse commonly reported parameters for all FPs in the database, and the comparison tab facilitates the collection of a set of FPs for the comparison of attributes complemented by sequence alignments (Supplementary Note 2). A basic REST API enables programmatic data retrieval (Supplementary Note 2). The interactive chart graphically displays relationships between fluorescence attributes across all FPs in the database, or in a filtered subset (Supplementary Fig. 2 and Supplementary Note 2). Integrated lineage trees depict the directed evolution of FPs, and users can search for proteins with specific mutations (Supplementary Fig. 3 and Supplementary Note 3).