Development priorities based on user feedback
- GROMACS: In biobb_gromacs we are working on including new features related to enhanced sampling simulations, taking advantage of the new integration of PLUMED in the GROMACS v2025.
Follow the developments in the corresponding GitHub repository (biobb_gromacs) - HADDOCK: In biobb_haddock we are expanding the set of building blocks to cover a higher part of the HADDOCK v3 features, now that the basic ones are implemented (see Implemented features section).
Follow the developments in the corresponding GitHub repository (biobb_haddock) - PMX: In biobb_pmx we are maintaining and updating the module to keep the pace with the new PMX implementations such as covalent or post-translational modifications.
Follow the developments in the corresponding GitHub repository (biobb_pmx)
- User-driven features and functionalities have been collected from users’ feedback and the 2024 BioExcel survey, and a set of new potential additions to the library has been defined, including new MD tools (OpenMM), enhanced sampling techniques (PLUMED, GROMACS AWH), structural conformation prediction and modeling (Rosetta, AlphaFold2), and QM methods (CP2K, Gaussian).
Follow the developments in our GitHub repositories (biobb)
- New building blocks are being developed in collaboration with the community: DNA-specific building blocks to implement a set of best practices in MD preparation, run, analysis and storage of nucleic acid simulations with the Ascona B-DNA consortium (hexABC project). Follow this development in the corresponding GitHub repository (biobb_dna)
New functionalities for the Virtual Screening module including popular tools in the field such as the AutoDock Vina fork Smina, the deep-learning docking tool Gnina, or the Gypsum-DL small molecule 3D predictor, in collaboration with the school of Pharmacy of the University of Eastern Finland. Follow this development in the corresponding GitHub repository (biobb_vs)
- New demonstration workflows to be deployed as Jupyter Notebooks or Google Colabs are under development, including a HADDOCK3 protein-protein docking, an AI–based feature extraction and pattern recognition from MD simulations using autoencoders, and a workflow to analyse and extract dynamic and flexibility properties from protein-membrane MD trajectories.
Follow these development in our GitHub workflows’ repositories (BioBB demonstration workflows)
- BioBB pre-exascale workflows to run massive calculations are being designed for a set of showcases in the BioExcel Centre of Excellence. Projects including coevolution-driven metadynamics simulations from ML-derived transition coordinates or AI-based insights on the functional annotation of sequence variants from MD simulations are being developed and will be tested in the CSC LUMI and BSC Marenostrum 5 supercomputers using the BSC PyCOMPSs workflow manager.
Follow these development in our GitHub HPC workflows’ repositories (BioBB HPC workflows)
Implemented features based on user feedback
- Following users’ requirements, a new biobb_haddock package has been developed, wrapping the new HADDOCK3 software and containing a collection of building blocks to compute information-driven flexible protein-protein docking.
This implementation can be found in the corresponding GitHub repository (biobb_haddock) - Following users’ requirements, the biobb_pmx module has been updated to the last PMX version, which accepts alchemical ligand modifications.
This implementation can be found in the corresponding GitHub repository (biobb_pmx)
- Following users’ requirements, a new biobb_pytorch package has been created, as a collection to train Machine Learning & Deep Learning models using the popular PyTorch Python library. The first building blocks implemented allow the training of an Auto-associative Neural Network (AANN) autoencoder that can be applied to reduce the dimensionality of MD data and analyze the dynamic properties of the system.
This implementation can be found in the corresponding GitHub repository (biobb_pytorch) - Following users’ requirements, a new biobb_mem package has been created, collecting a number of protein-membrane system-specific MD trajectories analysis and manipulation, using tools like MDAnalysis, FATSLiM and LiPyphilic.
This implementation can be found in the corresponding GitHub repository (biobb_mem) - Following users’ requirements, a new biobb_pdb_tools package has been created, wrapping a selection of functionalities from the HADDOCK’s team pdb_tools library, to easily manipulate and edit PDB files.
This implementation can be found in the corresponding GitHub repository (biobb_pdb_tools)
- In collaboration with the ELIXIR 3D-BioInfo structural community, a new module for the generation of protein conformational ensembles from 3D structures and the analysis of its molecular flexibility has been developed: biobb_flexdyn. The module includes a collection of tools created by members of the ELIXIR community.
This implementation can be found in the corresponding GitHub repository (biobb_flexdyn)
- A new demonstration workflow was created in collaboration with the ELIXIR 3D-BioInfo structural community: FAIR workflows to chart and characterize the conformational landscape of native proteins (Poster).
This implementation can be found in the corresponding GitHub repository (biobb_wf_flexdyn) - Following users’ requirements, the BioBB demonstration workflows are now available also from Docker containers (with new Docker images and Dockerfiles), and also Google Colabs. More information regarding the process followed to generate and use interactive Jupyter Notebooks and BioConda for FAIR and reproducible biomolecular simulation workflows can be found in the published paper.
These implementations can be found in the corresponding GitHub repository (biobb_workflows)
