IMP Code Roadmap

IMP's broad goal is to contribute to a comprehensive structural characterization of biomolecules ranging in size and complexity from small peptides to large macromolecular assemblies, by integrating data from diverse biochemical and biophysical experiments. IMP provides an open source C++ and Python toolbox for solving complex modeling problems, and a number of applications for tackling some common problems in a user-friendly way. IMP can also be used from the Chimera molecular modeling system, or via one of several web applications.

• extend IMP with a pytorch backend
  • create tensor-based representations
  • create restraints that can be differentiable
  • use builtin probabilistic programming samplers or generative AI sampler
  • create easy-to-use validation tools
• make IMP faster
  • CPUs
  • GPUs
• incorporate metamodeling
• support mixing and matching of multiple coupled representations
• formalize universal communication protocol for interfacing tools
• formalize community development and use of forward models and noise models
• incorporate Bayesian model validation
• using AI to reduce the technical complexity of using IMP
  • streamlining documentation
  • writing more examples and tests
• construct an argument to maximize the development of IMP versus other programs
• seek input from all IMP developers (including potential developers; IMP developers' workshop)
• maximize harmony between IMP and PDB-IHM
• reconsider the license to make IMP maximally useful to everyone
• expand the IMP distribution with other packages we develop (e.g. python-ihm, python-modelcif)