SBGrid Newsletter: March 2026 |
Dear Consortium Members and Affiliates, Our March member update includes a profile on SBGrid member Andrés Palencia, a reminder about our upcoming scipionTomo webinar, a software push with 16 updates and ten new titles, six new members to welcome, and updated documentation to help those new to SBGrid and or looking to fine tune the environment.
Our member tale for March features Andrés Palencia from the Institute for Advanced Biosciences, INSERM in Grenoble, France where his group is harnessing the matchmaking qualities of the LeuRS enzyme in efforts to advance a new class of drugs to tackle drug-resistant bacteria and parasites. Outside the laboratory, one might discover Palencia cycling the steep alpine peaks circling Grenoble or indulging a love of winemaking, nurtured in his family's vineyards in La Mancha, Spain. [Read More]
Join our software webinar on April 14th to hear from Jose Luis Vilas on Image processing made it simple with scipionTomo.. If you missed our March webinar with Josh Mitchell introducing Open Force Field, you can catch it on the SBGrid YouTube channel: https://www.youtube.com/user/SBGridTV To receive email reminders about upcoming webinars, please be sure to register for the series! Registration here: https://sbgrid.org/webinars/#register |
|
|
April 14: Jose Luis Vilas - Image processing made simple with scipionTomo May 12: Jarrett Johnson - Upcoming improvements to visualization and CryoEM support in PyMOL
June 9: Mart Last - Ais and Pom
Webinar registration and details |
SBGrid webinars are hosted with partial support from the NIH R25 Continuing Education for Structural Biology Mentors #GM151273, in collaboration with Co-PI Jamaine Davis of Meharry Medical College. |
|
|
This month's software push includes updates to AlphaPulldown, AreTomo, Coot, CCP4, CryoLithe, DIALS, Doppio, Java, MPicker, OpenFreeEenergy, Open Force Field, qfit, SAMtools, Schrodinger, Scipion, and Xmipp, along with 10 new applications: AFmassive, MemBrain-v2, Napari-SegSelect, MiCSPARC, MolSnapper, MolDiff, MassiveFold, Protein-quest, PXDesign, PXDesignBench
See Software Changes below for complete details.
Six new members joined in the month of March: Eddy Arnold from Rutgers University, Isaac Fianu from California Institute of Technology, Mamuka Kvaratskhelia from University of Colorado Anschutz, Ryan Nett and Nicholos Bellono from Harvard University, and Rebeccah Warmack from Yale University, Welcome to our newest members! |
Technical Notes from our Software Team |
Updated SBGrid Documentation for Installation and Daily Use
We recently updated the SBGrid wiki to include more complete guidance for new and existing users.
Our revised installation instructions provide clearer, step-by-step instructions for setting up SBGrid: https://docs.sbgrid.org/installation/installing_sbgrid/ Expanded Using SBGrid documentation provides more detailed guidance on working within the SBGrid environment: https://docs.sbgrid.org/environment/using_sbgrid/
We encourage all users to review these updates, whether installing SBGrid for the first time or looking to make better use of the environment in your daily work. Both resources are directly accessible from the sbgrid.org website. |
Member Publication Highlights |
Nearly 100 new member publications appeared in journals this month. You can find a complete listing on our website. |
Deposit your experimental datasets |
If you're currently preparing a manuscript, please remember that while you're making the PDB record deposit and publication submission, you can also preserve your primary experimental datasets with deposits to the SBGrid Data Bank: https://data.sbgrid.org/ |
SBGrid operations are primarily funded with member fees and grants, so we are grateful when you can acknowledge SBGrid in manuscripts, preprints, and related work where appropriate. We recommend the following language for inclusion in the methods section of your publications that report results obtained with SBGrid supported software: |
Software: Computational analyses were performed using the SBGrid software environment, which provides curated, version-controlled structural biology applications and reproducible execution environments across platforms (Herre et al., 2024).
Software and computational environment: All computational analyses were carried out using the SBGrid software environment, which provides curated, version-controlled access to structural biology software and manages dependencies to support reproducible research. Software execution was supported through the SBGrid Capsules framework, which standardizes application environments and workflows across platforms, as described in Herre et al. (2024).
|
SBGrid's eLife and Acta Crystallographica D publications received two new citations since our last reporting, from these SBGrid members and contributing developers: Christopher Garcia from Stanford University School of Medicine in Science: Overcoming T cell tolerance to tumor self-antigens through catch-bond engineering. https://www.science.org/doi/abs/10.1126/science.adx3162
Dhirendra Simanshu and Anna Maciag from the Frederick National Laboratory for Cancer Research in Cancer Discovery: Discovery of BBO-11818, a Potent and Selective Noncovalent Inhibitor of (ON) and (OFF) KRAS with Activity against Multiple Oncogenic Mutants.
https://aacrjournals.org/cancerdiscovery/article/doi/10.1158/2159-8290.CD-25-1280/774990 |
AFmassive v1.1.6 is a modified AlphaFold implementation for massive conformational sampling, derived from AlphaFold v2.3.2 and designed to integrate with MassiveFold for parallelized prediction workflows. It adds inference-time diversity controls including dropout in Evoformer and the structure module, configurable dropout-rate profiles, expanded recycle and early-stopping settings, model-version selection across AlphaFold 2.1–2.3 multimer parameter sets, and prediction-range controls for distributed runs. The software targets high-throughput protein structure prediction campaigns where broad sampling and scalable execution are needed.
https://github.com/GBLille/AFmassive/
MemBrain-v2 0.0.1 includes MemBrain-seg, MemBrain-pick, MemBrain-stats, and several Napari tools. https://github.com/CellArchLab/MemBrain-v2 Napari-SegSelect 0.1.3 is a plugin for Napari to select a connected component from a Membrain-seg segmentation.
https://github.com/bwmr/napari-segselect
MiCSPARC 1.0 is a CryoSPARC-based cryo-EM pipeline for microtubule image processing that combines automated filament tracing, particle picking, fast 3D refinement, protofilament-number assignment, alignment, register correction, and seam search to determine structures of decorated and undecorated microtubules. It includes Python utilities for filament extrapolation, reference generation, protofilament assignment, and angular unification, with a conda environment targeting cryosparc-tools 4.7.1. The workflow is designed for end-to-end microtubule reconstruction in CryoSPARC, from curated exposures and template generation through high-resolution protofilament and full microtubule refinement.
https://github.com/wieczoreklab/MiCSPARC
MolSnapper 20250225 is a tool for conditioning diffusion models to generate 3D drug-like molecules, built on the MolDiff codebase. It integrates pretrained models and provides a streamlined workflow for sampling molecules from defined pockets, making it suitable for structure-based drug design applications. The software supports various data processing tasks, including dataset preparation and evaluation of generated molecules.
https://github.com/oxpig/MolSnapper/
MolDiff 20260225 is a diffusion framework designed for generating 3D drug-like molecules, achieving a sampling success rate of over 99%. This tool addresses the atom-bond inconsistency problem in 3D molecule diffusion generation and serves as a backbone for applications such as pocket-based generation and linker generation.
https://github.com/pengxingang/MolDiff/
MassiveFold 1.6.0 is a tool designed to expand structure prediction sampling by optimizing the parallelization of AlphaFold-based predictions across GPU clusters. It automatically splits large prediction runs into batches distributed across multiple GPUs, with alignment computations performed on CPUs, and gathers all results into a single ranked output with comprehensive visualization plots. This tool supports AFmassive, ColabFold, and AlphaFold3 as inference engines and is optimized for SLURM workload managers, while remaining usable for sequential execution on single-GPU servers.
https://github.com/GBLille/MassiveFold
protein-quest 1.1.2 is a Python package to search/retrieve/filter proteins and protein structures. It includes Uniprot Sparql endpoint to search for proteins and their measured or predicted 3D structures, Uniprot taxonomy to search for taxonomy, QuickGO to search for Gene Ontology terms, gemmi to work with macromolecular models, dask-distributed to compute in parallel, and rocrate-action-recorder for provenance tracking.
https://www.bonvinlab.org/protein-quest/
PXDesign 20260305 is a de novo protein-binder design suite combining a diffusion generator (PXDesign-d) with Protenix and AF2-IG confidence models for candidate selection and ranking. It supports YAML-defined target preparation, MSA-assisted extended runs, preview and generation-only modes, and output filtering with AF2-IG and Protenix success criteria to prioritize binders for wet-lab validation. https://github.com/bytedance/PXDesign/
PXDesignBench 0.1.1 is a protein design evaluation suite for monomer and binder tasks that standardizes sequence generation, structure consistency assessment, and downstream metrics across integrated tools including ProteinMPNN, ESMFold, AlphaFold2, and Protenix. It provides task-specific pipelines, wrappers for external models, and configurable JSON- or directory-based inputs for batch evaluation, bias-controlled ProteinMPNN sampling, and multi-GPU or distributed execution. Outputs include confidence and self-consistency metrics, designability and shape-based structural statistics, and summary CSVs for comparative analysis and post-processing.
https://github.com/bytedance/PXDesignBench |
AlphaPulldown 2.2.0 is a correctness and stability release that fixes AlphaFold3 MSA order (using AlphaFold2 features with the AlphaFold3 backend could lead to omitting paired MSAs) and an issue with output dir of the AlphaFold3.
https://github.com/KosinskiLab/AlphaPulldown/releases/tag/2.2.
AlphaPulldown 2.1.9 fixes chain identity (asym_id) and adds proper AF3 homooligomer copy support.
https://github.com/KosinskiLab/AlphaPulldown/releases/tag/2.1.9
AreTomo 2.2.8 includes -CorrCTF 3 for local CTF multiplication in template matching, improved thickness measurement with WBP, a fix to forward projection between ProjAlign and Recon, includes new GGenRandoms.cu and GFillEmpty2D.cu utilities to fill empty regions/pixels, modified reconstruction code to better handle empty voxels and positivity constraints, removal of an unreliable tilt-axis refinement, and the addition of CUDA 13 support. In this release the developers also fixed forward-projection issues when working with large negative values and incorrect diff projections, vertical striations in SART, an indexing error in GExtractTile.cu, and an issue in CAreTomo3Json.cpp.
https://github.com/czimaginginstitute/AreTomo3/blob/main/Readme.txt
Coot 1.1.20 focuses on expanding the software's functional depth through new chapi Pythonic API functions, specifically adding get_mmrrcc() and get_residue_pucker_info() for more detailed model analysis. This version introduces several user-centric features, including dedicated key-bindings for Triple-Refine AA, Undo Symmetry View, and Keyboard Mutate, alongside a new "Unhappy Atom" markup to aid in validation. Practical interface improvements include open dialogs that now track the most recently used directory and a bug fix that ensures the "Get monomer" frame is correctly undisplayed upon activation.
https://github.com/pemsley/coot/releases
CCP4 9.0.014 delivers a newer version of Crank2 (v2.0.317) and introduces the Slice n' Dice interface, alongside specific bug fixes for the Arcimboldo interface. For molecular weight calculations, the ASU Content task received revisions and fixes, while Refmac was expanded with a new option for neutron diffraction. This release also aligns the software's documentation theme with cloud-based standards and resolves a reported connectivity issue involving OpenVPN.
https://www.ccp4.ac.uk/ccp4-9-0-updates/#Update_90014
CryoLithe v1.0 includes the code with the original interface and links to the model. https://github.com/swing-research/CryoLithe/releases/tag/v1.0
DIALS 3.27.0 includes significant enhancements, such as new parameters for improved indexing solution filtering, support for LCLS-II file formats, and the introduction of a time-of-flight data processing tool. This version also features a utility for checking space group symmetry, various output format improvements, and numerous bug fixes aimed at enhancing performance and stability. https://github.com/dials/dials/releases/tag/v3.27.0
Doppio 1.5.0 key new features include an option to use Doppio in multiple projects simultaneously, improvements to job searching and results displays, and new SSL options for Doppio server. This version also fixes an issue with passing the correct Topaz executable name to both RELION 4 and 5 and adds the ability to continue a MotionCorr job to regenerate output micrographs from previously-calculated shifts, along with other minor bug fixes and updates.
https://www.ccpem.ac.uk/docs/doppio/release_notes.html
Java 1.8.0_481 is a minor version update to Java8 for security and bug patches. https://www.oracle.com/java/technologies/downloads/#java8-mac
MPicker 1.3.0 adds Mpicker_merge.py to merge two slices in a flattened tomogram with different color channels, Mpicker_autoextract.py to automatically extract and flatten surfaces from membrane segmentation, a function in the GUI to flatten all selected surfaces in batch, and an Mpicker_class2d.py as a standalone program to do 2D classification.
https://github.com/yanxf1/MPicker_v1.3/blob/main/releaseNote.md#mpicker-version-130-2025-12-18
OpenFreeEenergy 1.9.1 includes several important fixes and improvements, such as resolving an API break in tests, ensuring safe closure of reporters and GPU contexts to prevent UnboundLocalErrors, and correcting the logging configuration import in the command line interface, in addition to enhancing stability and reliability.
https://github.com/OpenFreeEnergy/openfe/releases/tag/v1.9.
Open Force Field Toolkit 0.18.0 includes minor API changes, performance improvements, and bug fixes. The internal method Molecule.are_isomorphic.to_networkx was removed from the public API to allow future optimizations; Molecule.are_isomorphic and is_isomorphic_with performance is improved, which can speed up interchange creation and export; and mappings for symmetric atoms may differ but remain chemically equivalent. A bug causing unnecessary conformer coordinate lookups for large molecules was also fixed.
https://docs.openforcefield.org/projects/toolkit/en/stable/releasehistory.html#id1
qfit 2025.3 includes significant improvements, such as the introduction of ensemble to multiconformer conversion scripts and enhanced occupancy redistribution based on rotamer states. This release also addresses multiple bug fixes, including issues with residue selection and refinement scripts, ensuring a more robust modeling experience for users. https://github.com/ExcitedStates/qfit-3.0/releases/tag/2025.3
SAMtools 1.23.1 primarily focuses on critical stability fixes, addressing a regression in the depth command that caused incorrect results and fixing several memory-related crashes. The update also improves efficiency by optimizing CRAM and VCF processing and ensuring better compatibility with recent HTSlib changes.
https://github.com/samtools/samtools/releases
Schrodinger 2026-1 includes a redesigned Surface Manager for enhanced visualization control, persistent geometric measurements for structural comparisons, and standalone density map import to streamline Cryo-EM workflows. This update also introduces a context-aware Maestro Assistant in open beta, facilitating seamless interaction between documentation and direct actions. https://www.schrodinger.com/life-science/download/release-notes/
Scipion 3.11.6 updates all the Tomography plugins including AreTomo, cryoCARE, Deep Finder, Dynamo, emantomo, fidder, GAPSTOP(TM), IMOD, MemBrain, MotionCorr, NovaCTF, Reliontomo, Sphire (crYOLO and Janni), Tardis, TomSsegMemTV and Warp. Also included is a major update of the Xmipp plugin, which now runs version 5.0.0-betagal, and the addition of Gautomatch, Repic, Cryoasses, CryoDrgn and Scipion-for-facilities plugins to Scipion in SBGrid.
Xmipp 5.0.0-betagal includes a new 2D classification algorithm for SPA (alignPCA_2D program - an alignment method applied for 2D classification using PCA and Euclidean distance), and full CUDA 11 → 13 compatibility, with more GPU flexibility and smoother setups. This release also includes bugfixes and general improvements. https://i2pc.github.io/docs/Releases/Releases-xmipp-program/index.html |
Please note that not all software applications are available to every SBGrid member type. If you see an application that you would like to use, but is not included in your software tree, please contact us to find out what options are available for access. Learn more about SBGrid - Consortium website: https://sbgrid.org - Supported software: https://sbgrid.org/software/
- Report software bugs: sbgrid.org/bugs
This newsletter is sent to you because you are a member or affiliate of the SBGrid Consortium. |
|
|
|
