SASBDB: Towards an automatically curated and validated repository for biological scattering data

Abstract

Small-angle scattering (SAS) of X-rays and neutrons is a fundamental tool to study the nanostructural properties, and in particular, biological macromolecules in solution. In structural biology, SAS recently transformed from a specialization into a general technique leading to a dramatic increase in the number of publications reporting structural models. The growing amount of data recorded and published has led to an urgent need for a global SAS repository that includes both primary data and models. In response to this, a small-angle scattering biological data bank (SASBDB) was designed in 2014 and is available for public access at www.sasbdb.org. SASBDB is a comprehensive, free and searchable repository of SAS experimental data and models deposited together with the relevant experimental conditions, sample details and instrument characteristics. SASBDB is rapidly growing, and presently has over 1,000 entries containing more than 1,600 models. We describe here the overall organization and procedures of SASBDB paying most attention to user-relevant information during submission. Perspectives of further developments, in particular, with OneDep system of the Protein Data Bank, and also widening of SASBDB including new types of data/models are discussed.

Figure 1

Large scale synchrotron radiation and neutron facilities contributing to SASBDB

Figure 2

SASBDB contents by model type

Figure 3

Representation of a SASBDB entry (based on https://www.sasbdb.org/data/SASDDF6/). (a) Title of the publication (or project in case of unpublished data). (b) List of authors (the main contributor is highlighted in bold), journal reference and a link to Europe PubMed Central. (c) SASBDB code and the title of the entry. (d) Names of the macromolecules that were measured. (e) Experimental scattering data (logarithmic plot). (f, g) Experimental data range percentile ranks. (h) Experimental noise percentile rank. (i) Fit to the experimental data. (j) Error‐weighted residual difference plot. (k) Goodness‐of‐fit percentile rank. (l) Experimental details. (m) Entry tag. (n) Drop‐down list of files available for download. (o) Comparison between experimental and expected molecular weights. (p) Guinier plot with the linear fit and the estimated values of the forward and scattering I(0) and the radius of gyration R _g. (q) Dimensionless Kratky plot. (r) Pair distance distribution function p(r) and the maximum intraparticle distance D_max. (s) p(r) goodness‐of‐fit to data percentile rank. (t) Histogram of the R _g distribution of the generated pool (grey bars) compared to the final ensemble (blue bars), if present. (u) Associated models, for example, the most representative models from EOM ensemble. (v) Biological details of each measured macromolecule, FASTA sequence and link to UniProt

Figure 4

Histogram of χ² distribution of p(r) fits with CorMap test p‐values >.01 (64% of all data sets). The green bars represent 42% of entries with expected 0.8 < χ ² < 1.2. Inset: the ratio of acceptable p(r) fits (p‐values >.01, blue) and fits with systematic deviations (p‐values <.01, red) according to the CorMap test