New features in ATSAS-4, a program suite for small-angle scattering data analysis

This article is part of a collection of articles related to the 19th International Small-Angle Scattering Conference (SAS2024) in Taipei, Taiwan.

The ATSAS release 4 series introduces substantial architectural updates and enhanced user-friendly tools for analyzing biological macromolecules and nanostructured objects. In particular, highly requested project support for graphical user interfaces (GUIs) has been implemented. Here the state of the main window of each GUI application can be saved as a project and then restored to continue or extend analyses. All GUI applications now share a common framework, enabling various plot annotations and export as figures or text for easy data exchange with other plotting applications. The main interactive data processing and interpretation program, PRIMUS, has been significantly expanded with new features and integrated analysis modules. A new GUI for the program PEAK allows for multiple fitting functions, including Gaussian, Lorentzian and pseudo-Voigt. The online chromatography analysis program CHROMIXS supports multiple frame selections and subtractions, along with a newly included EFAMIX module for resolving partially overlapping components. The new ATSAS application LAUNCHER provides user-friendly wizards to configure and run command-line applications. Wherever feasible, operations like file reading and calculations have been parallelized to take advantage of available local compute resources. Command-line interfaces of many applications have been updated, e.g. with renaming and unification of common options. Improved readers/writers are implemented for Protein Data Bank and macromolecular CIF (mmCIF) file handling. The program MIXTURES now combines the functionalities of BILMIX, LIPMIX and MIXTURE, offering simplified and more versatile analyses for polydisperse systems of different shapes, also with single- or multi-layered organization. Installer packages of ATSAS are available for the latest versions of Windows, frequently used Linux distributions and recent macOS releases, both for Intel and ARM architectures, and are free for academic use.

1. Introduction

Small-angle scattering (SAS) is a powerful technique widely employed in structural biology for the study of biological macromolecules in solution (proteins, nucleic acids and large complexes) and numerous polymeric and nano­structured objects, e.g. lipid nanoparticles. Its primary strength lies in the ability to investigate macromolecular conformational changes, flexibility and interactions under near-physiological conditions. By providing information about the overall shape, size and structural organization of molecules in solution, SAS offers critical insights not only into individual biomolecules but also into larger complexes and mixtures that may be challenging to analyze using high-resolution techniques. SAS has become a key component in integrative structural biology, where it complements data from high-resolution methods, such as X-ray crystallography and cryo-electron microscopy, to offer a more complete understanding of macromolecular structures and their dynamic behavior (Jeffries et al., 2021). Consequently, much work is underway aiming at standardization, quality, reproducibility and validation of SAS data and models (Trewhella et al., 2022; Trewhella, 2023).

The ATSAS software package is integral to the efficient processing, analysis and interpretation of small-angle X-ray scattering (SAXS) and small-angle neutron scattering data. It includes a wide array of tools for tasks such as primary data processing, ab initio shape reconstruction, and the hybrid analysis of monodisperse systems, mixtures, and flexible or partially disordered systems (Manalastas-Cantos et al., 2021). Over the past 30 years, ATSAS has continuously evolved to meet the increasing complexity of SAS data and the expanding needs of the scientific community. Early versions were centered around basic operations like data reduction and simple model fitting (Konarev et al., 2006), while later iterations introduced more advanced tools for hybrid modeling, integrating SAS data with atomic structures from X-ray crystallography and nuclear magnetic resonance (Petoukhov et al., 2012; Franke et al., 2017). The software now supports the analysis of flexibility, mixtures and time-resolved experiments, making it broadly applicable for modern SAS studies. Time-resolved experiments can also be analyzed, which, with modern equipment, may be conducted not only on high-brilliance synchrotrons but also on laboratory X-ray sources (Rasmussen et al., 2023). Additionally, modules were added to support the analysis of SAS data from soft matter systems and nanostructured materials.

The first applications that eventually became part of ATSAS, e.g. GNOM (Svergun, 1992), were developed, as customary at the time, exclusively as command-line tools. Most ATSAS applications are still available as command-line applications only. More recent software projects like SASView (https://www.sasview.org), Scatter (https://github.com/rambor/scatterIV), McSAS (https://github.com/BAMresearch/McSAS3), SAXSutilities (Sztucki, 2021) and BioXTAS RAW (Hopkins, 2024) started out with graphical user interfaces (GUIs). These tools were more easily accessible by younger researchers accustomed to the GUI workflow who may consider command-line applications less intuitive. That said, Konarev et al. (2003) already initiated this shift by introducing a GUI for SAS applications, followed by the first full release of ATSAS (Konarev et al., 2006). However, these GUIs were limited to Windows PCs. This limitation was addressed later by adopting a cross-platform GUI framework, extending graphical access to macOS and Linux users while also improving the usability of key tools like PRIMUS (Franke et al., 2017). Nevertheless, the majority of ATSAS applications remained available only via the command line, posing a barrier for some users.

To mitigate, ATSAS Online (Petoukhov et al., 2012) and the ATSAS Portal on the grid (Wassenaar et al., 2012) offered web-based alternatives to traditional GUIs. Hosted on a small-scale cluster at EMBL, Hamburg Unit, ATSAS Online provided access to essential applications through a simple web interface. With the development of an ATSAS Portal for the grid, these initial computational resources were extended by integrating a large number of global compute nodes. This advancement enabled researchers to handle computationally demanding tasks, such as ab initio and hybrid modeling as well as ensemble analysis, on a much larger scale, without the need for a command line. However, these remote resources may be difficult to utilize by researchers working on confidential projects due to the openness of the architecture, potentially allowing access to the data and its analysis by third parties.

The ATSAS-4 release series introduces new graphical interfaces to further improve accessibility, including enhancements to cross-platform applications like PRIMUS and CHROMIXS and the addition of new GUIs such as PEAK and LAUNCHER. New programs are included, e.g. EFAMIX for advanced analysis of mixtures. Numerous modules have been optimized for computational efficiency, read/write performance and interface organization.

This article exclusively focuses on the new developments in ATSAS-4 and is not intended as an exhaustive overview of all components of the ATSAS package. The interested reader is referred to previous ATSAS publications for overviews provided at the time and the detailed publications that are available for most programs.

2. GUIs

Multiple authors working on different applications may differently implement similar functionalities. In previous ATSAS releases, for example, CHROMIXS and PRIMUS (Franke et al., 2017) were built using the same GUI toolkit and plotting library, but provided significantly different interfaces. Specifically, features and workflows in one application did not seamlessly translate to the other. For ATSAS-4, we addressed this issue by developing a unified data model and view for SAS plots within the Qt model/view architecture (https://doc.qt.io/qt-6/model-view-programming.html). We then re-implemented all plot windows and wizards in ATSAS to utilize this common framework. As a result, every plot in every application now shares consistent features, allowing workflows learned in one ATSAS application to be applied across others. Notable features of this framework include:

(i) Parallelized reading of data files for improved performance.

(ii) Capability of plotting millions of data points simultaneously.

(iii) Full configurability of all plot aspects (axes, data series, annotations, legend).

(iv) Freely placeable horizontal/vertical lines, text annotations and images.

(v) Export options for figures in various formats (PNG, BMP, JPEG, TIFF, SVG or PDF).

(vi) Export of all plot data in current coordinates to CSV for easy import into other plotting applications.

The common data model also made it possible to introduce one of the most user-requested features across all applications: project support. Before ATSAS-4, GUI applications either did not retain any information from previous sessions or, at best, maintained a list of recently opened files. With project support, users can save the state of an application, e.g. currently open files, plot configurations etc., and restore it later. For instance, in PRIMUS, this includes all data with their respective selections and the presently displayed graphical window configuration. However, previous analysis results, such as the radius of gyration (R_g) for each data file, are not yet stored.

In addition to the plot model/view architecture, we also enhanced the wizard framework first introduced in PRIMUS (Franke et al., 2017). Here, a wizard is a step-by-step guide that helps users to accomplish a specific task by breaking it down into smaller, manageable stages through dedicated wizard pages. Each page collects information or sets preferences, thus making complex processes more approachable and intuitive for users. Internally, we improved page independence and reusability, which streamlined the data flow across wizards. The most notable change for users is that, starting with ATSAS-4, wizards can execute tasks in parallel. For example, after configuring ab initio modeling, the wizard may run multiple DAMMIF (Franke & Svergun, 2009) tasks simultaneously, depending on the number of available system cores, hence significantly reducing overall runtime. Various wizards implement this feature in PRIMUS, CHROMIXS and the new LAUNCHER application.

2.1. LAUNCHER

One common challenge for new users of ATSAS is navigating through the wide array of available tools and options. To address this, the `ATSAS application launcher' was developed as a user-friendly graphical interface designed to simplify access to complex command-line applications (Fig. 1). By providing a collection of intuitive wizards, LAUNCHER streamlines the configuration and application of these tools, reducing the learning curve for users.

Figure 1 GUI of the ATSAS application LAUNCHER with a running DAMMIF wizard. LAUNCHER offers a convenient way to start other GUI applications and also a command-line console. Furthermore, it provides access to a variety of application wizards. Here the wizard for the bead modeling application DAMMIF (Franke & Svergun, 2009) is shown. Although six repetitions of DAMMIF were configured, given the hardware limitations of a virtual machine, only two are running in parallel (alternating yellow/dark yellow indicators), two are already completed (green), and two more are pending (white). It is possible to start multiple wizards in parallel; CPU resources will then be shared among them.

LAUNCHER offers wizards for the most popular tasks such as ab initio modeling, hybrid modeling, model manipulations, and mixture analysis and flexible systems. Each wizard operates independently, allowing users to run multiple wizards simultaneously, with all tasks managed by a shared internal queue. Much like a cluster task manager that distributes jobs across multiple nodes, LAUNCHER allocates system resources, running only as many tasks as available CPU cores allow to prevent the system overload. This design is functionally similar to submitting tasks through an ATSAS Online interface but operates locally and securely. Each wizard with three or more pages shows an outline of the required configuration, indicating the completed steps. Once the wizard tasks are submitted, the final configuration is saved to a .launcher file in the working directory. This configuration can be used as a starting point for a new task by opening it with the LAUNCHER application and then applying any required changes before resubmitting to the internal task queue. LAUNCHER is particularly beneficial e.g. for running multiple jobs overnight or when handling complex analyses.

In addition to the wizards, LAUNCHER provides quick access to other graphical ATSAS applications and includes a shortcut for opening a pre-configured console window, granting users a convenient way to utilize applications that are not yet available as wizards.

2.2. PRIMUS

PRIMUS (Konarev et al., 2003) has been a core component of the ATSAS suite since its first release (Konarev et al., 2006), specifically designed for the analysis of SAS data from macromolecular solutions. It remains one of the most versatile tools for SAS data analysis and, after porting to the common framework developed for ATSAS-4, PRIMUS is now capable of handling large datasets with thousands of frames and millions of data points. This makes it suitable for both single-dataset analysis and large-scale data processing tasks.

PRIMUS provides access to a comprehensive range of tools for basic data reduction and advanced analysis. These include modules for comparing and averaging data frames, background subtraction, and scaling and merging data interactively, allowing users to fine-tune their datasets before further analysis. The software also integrates numerous analysis wizards, enabling users to determine key structural parameters such as the R_g, molecular weight (MW), maximum particle diameter (D_max) and Porod volume (V_p). In addition to data manipulation, PRIMUS supports more advanced analyses through wizards for ab initio shape reconstruction and hybrid modeling.

Like other GUI applications in the ATSAS suite, PRIMUS allows users to save their work in project files, preserving the state of the analysis, plot configurations and data selections.

2.3. CHROMIXS

CHROMIXS (Panjkovich & Svergun, 2018) is designed for convenient analysis of size-exclusion or ion exchange chromatography coupled with SAXS (SEC-SAXS, IEX-SAXS) data. It automates the reduction and interpretation of large datasets generated during SEC-SAXS experiments by offering a user-friendly graphical interface that visualizes the data as averaged scattering intensity versus elution time (Fig. 2). Users may manually select the sample and buffer regions or let the program automatically detect these regions using a built-in peak-finding algorithm.

Figure 2 GUI for SEC-SAXS data analysis program CHROMIXS, here with data from the Small Angle Scattering Biological Data Bank: SASDF99 (BSA). The intensity trace together with multiple selections is shown in the left window. The top right-hand table lists the three selections: one for the monomer peak (around frame #1600), one for the dimer peak (around frame #1350) and one for the Rg calculations in a wider range (frames #1300–#1700). The plot at the center right shows all selected data frames with the current selection, here the monomer peak, shown in the corresponding colors. The plot at the lower right shows the two subtracted datasets, one for the monomer (blue), the other for the dimer (brown).

In ATSAS-4, besides the plot-related improvements mentioned above, CHROMIXS was upgraded to allow for multiple independent selections, representing different peaks or alternative selections of the same peak with varying buffer regions. This new feature, for example, allows for direct comparison of different subtractions within the interface. For each subtraction, CHROMIXS calculates key structural parameters, such as the R_g, MW and D_max, providing immediate feedback. Multi-peak systems may be further analyzed via evolving factor analysis, provided by EFAMIX (described below).

Opened data, frame selections and plot configuration may be saved in a project file, enabling users to resume their work later. Additionally, a `Package project' feature is available, which creates a .zip archive containing all data and a corresponding project file, making it easier to share and collaborate on large datasets.

2.4. PEAK

PEAK (Konarev et al., 2003) is designed to analyze scattering patterns of partially ordered systems. In ATSAS-4, the original Windows version of PEAK has been reimplemented and extended to use the Qt graphical software. Users interactively identify peaks in the scattering data by selecting them directly on the plot, after which the program fits the selected peaks using one of three configurable fit functions (Fig. 3): (i) Gaussian, (ii) Lorentzian or (iii) a Gaussian–Lorentzian mixture (also known as pseudo-Voigt). The software also allows for a two-parameter background subtraction to improve the accuracy of the peak fitting.

Figure 3 GUI of PEAK. Shown is the scattering from gold nanoparticles (data included in the ATSAS distribution package). Two peaks have been selected and the corresponding results are shown in the table.

PEAK calculates several important structural properties from the scattering data, including d spacing, which corresponds to the distance between repeating units in the structure, as well as long-range order dimension L, interaction radius R_m and disorder level Δ/d. These parameters provide quantitative insights into the structural ordering in the sample. The results are visually represented as fitted curves on the plot, and all calculated values are listed in an accompanying table. When saved as a project, the software preserves the peak fits, structural parameters and plot annotations, such that the work can be seamlessly resumed as needed.

3. Polydisperse systems

Besides improvements in the GUIs, ATSAS-4 also provides new and updated applications for the analysis of polydisperse systems.

3.1. EFAMIX

EFAMIX (Konarev et al., 2022) is designed to work with SEC-SAXS or IEX-SAXS data. SEC-SAXS became a very popular tool in SAXS studies and multiple approaches were developed for extraction of pure component patterns from poorly resolved SEC-SAXS data (Brookes et al., 2016). The EFAMIX program is able to analyze data from partially overlapping elution peaks, a common challenge when multiple components are not fully separated on the chromatographic column. By using evolving factor analysis (EFA) (Maeder, 1987), EFAMIX can restore individual scattering and concentration profiles from such mixtures, providing accurate data decomposition with minimal user intervention.

One of the core features of EFAMIX is its ability to handle complex, multicomponent systems. It employs a model-free approach using singular value decomposition and rotation matrices to separate overlapping scattering profiles. This method is particularly useful for biological macromolecules, which are often present in mixtures or in partially separated complexes even after chromatography purification.

Given the expected number of components, the frame range for analysis and the angular range for SAXS data processing, EFAMIX automatically predicts the concentration windows for each component and provides both the concentration profiles and the restored scattering curves of the components. These results can then be compared with theoretical scattering models or experimental data for further analysis.

EFAMIX can be used from the command line or through the `Evolving factor analysis' wizard from CHROMIXS, based on the frame selection option of the latter.

3.2. MIXTURES

MIXTURES combines the functionality of three independent tools: MIXTURE (Konarev et al., 2003), LIPMIX (Konarev et al., 2021) and BILMIX (Konarev et al., 2020). This integration enhances the analysis of complex systems composed of multiple components, such as polydisperse mixtures, lipid bilayers and vesicles, by offering a unified instrument for SAS data analysis.

The software models multicomponent systems with simple and advanced geometries. Building upon the original capabilities of MIXTURE, the program can handle various shapes, such as spheres, ellipsoids, cylinders and core–shell particles, incorporating parameters like size, shape and polydispersity. It also facilitates the quantitative analysis of systems where multiple scattering components coexist with each other, allowing users to account for interparticle interactions and size distributions in complex mixtures.

By incorporating tools from LIPMIX and BILMIX, MIXTURES enables the analysis of bilayered systems using the methodology first described by Pabst et al. (2000). This enables the restoration of the electron-density profiles of lipid bilayers, accounting for their polydispersity and multilamellar organization. These features are crucial for studying vesicle-based systems, such as liposomes, nanodiscs or biological membranes.

The software supports SAXS data fitting across a broad angular range, which is particularly beneficial for analyzing lipid mixtures accounting for both form factor and Bragg peak scattering portions. These analyses yield bilayer electron densities, vesicle size distributions and other structural parameters within a single modeling step, rather than relying on separate fitting procedures.

Designed with flexibility in mind, the software supports automated batch processing and parallel computations. This makes it particularly well suited for high-throughput experiments and complex datasets where simultaneous modeling of multiple components is required. This consolidation of modeling approaches into a single program offers streamlined workflows and expanded capabilities for multicomponent systems analysis within ATSAS-4.

Users can access MIXTURES through the command line or via the `Mixtures' wizard in LAUNCHER. Wizards are also implemented for other programs dealing with mixtures and flexible systems, e.g. OLIGOMER (Konarev et al., 2003) and EOM (Bernadó et al., 2007; Tria et al., 2015).

4. Technical updates

In addition to functionality expansions and improvements in GUIs, other, more technical, issues and requests from users have been addressed.

4.1. CRYSOL

Starting with ATSAS 3.1, CRYSOL uses the Chemical Component Dictionary (https://www.wwpdb.org/data/ccd) to determine the number of implicit hydrogens assigned to any named atom. Since the dictionary file, components.cif, is very large, parsing its contents takes several seconds – often longer than the actual scattering calculation for small mol­ecules. This slowed down CRYSOL computations compared with the previous implementations. As of ATSAS 4.1, the monolithic database is divided into individual component files, allowing CRYSOL to parse only the required files for a given model, thus significantly improving performance for small models.

4.2. Reading/writing atomic coordinates

Many applications of the ATSAS suite require atomic coordinates as input or provide them as output. Inputs classically come in Protein Data Bank (PDB) format or, today, in macromolecular cif (mmCIF) format (Adams et al., 2019). In the 3.x series of ATSAS, read and write support for the mmCIF format was added and mmCIF was introduced as the default output format for models in all applications. Due to multiple reports of disrupted workflows caused by third-party software not yet being able to read mmCIF, we introduced the --output-format option to all relevant applications to have either PDB or mmCIF output format, so that users may select the required output format for their workflows.

In addition, PDB and mmCIF parsing was improved significantly and is now able to report on a multitude of issues, for example misaligned columns in PDB files, which used to be hard to diagnose in the past.

4.3. Signed packages

At the time of writing, downloading software packages from untrusted sources from the internet is discouraged; hence Windows and macOS displayed security warnings prior to running downloaded content on a system. BIOSAXS GmbH has invested in trusted certificates to sign the ATSAS package before its release. Signing ATSAS ensures that users are not unintentionally downloading malicious versions from untrustworthy sources, reducing the risk of malware infections or unauthorized access to systems. For researchers and organizations handling sensitive data, this safeguard is especially critical, as it helps maintain the security and integrity of both the software and the data processed by it.

4.4. ATSAS licenses

As of ATSAS-4, a license file is required to run ATSAS applications. Users of ATSAS that are affiliated with an academic or educational institute can request a free academic license from BIOSAXS GmbH by submitting an institutional email address (see https://biosaxs.com/download.php).

4.5. ATSAS at GitHub

ATSAS development has transitioned to Git, with all repositories now hosted on GitHub. This move enhances version control, allowing developers to track changes and collaborate more efficiently through branch management and concurrent development. GitHub further improves the process by offering secure, centralized repositories with granular access controls. It also enables continuous integration, automated testing and peer reviews, ensuring better transparency and code quality while streamlining the overall development workflow.

For registered ATSAS users, binaries of the latest releases are available for download from a dedicated community repository at GitHub: https://github.com/biosaxs-com/atsas-community/releases/. Users are encouraged to use the associated issue tracker (https://github.com/biosaxs-com/atsas-community/issues) for reports to the ATSAS developers and discussion forum (https://github.com/biosaxs-com/atsas-community/discussions) for general questions and user-to-user communication. Please note that the previously advertised (Manalastas-Cantos et al., 2021) user forum at saxier.org has been closed. However, all discussions previously found at saxier.org have been imported to GitHub Discussions and are preserved there.

5. Conclusion

The ATSAS-4 release series provides numerous new and enhanced features for the analysis of SAS data from biomacromolecular solutions and nanostructured systems. Many existing applications have been made more accessible and overall easier to use. The core component, PRIMUS, along with other GUI-based programs was rewritten in the Qt model/view architecture such that every application now shares consistent features. A new application, LAUNCHER, serves as a springboard to rapidly and conveniently access many ATSAS analysis programs, and the built-in task manager functionality helps to leverage local compute power. CHROMIXS now offers multiple frame selections and a wide range of options for SEC-SAXS data analysis including data processing by EFAMIX to handle partially overlapping elution peaks with minimal user intervention. PEAK provides convenient means for the analysis of partially ordered systems, while MIXTURES allows for comprehensive analysis of complex polydisperse systems containing multiple components. Further technical enhancements, integration and addition of new functionalities into ATSAS programs is under continuous development by BIOSAXS GmbH.

Licenses for ATSAS-4 are free of charge for academic users. Installers for Windows, macOS and multiple Linux distributions are available at the ATSAS Community repository at GitHub: https://github.com/biosaxs-com/atsas-community/.