*              PPSN 2014

September 13-17, 2014 - Ljubljana, Slovenia



13th International Conference on Parallel Problem Solving From Nature



Workshop “Natural computing for protein structure prediction”







Independent of its starting conformation, a protein in its natural environment folds into a unique three dimensional structure, the native structure. Understanding the native structure of a protein is crucial, as the structure can provide insight into the functional roles of a protein and the specific mechanisms of its biological function. The experimental determination of the native conformation (using X-ray crystallography or NMR spectroscopy) is difficult and time-consuming. As a result, the output of experimentally determined protein structures lags behind the output of protein sequences, and the computational prediction of protein structure remains a “holy grail” of computational biology.

Work related to the prediction of protein structure comprises the prediction of secondary and tertiary structure. In the case of secondary structure prediction, machine learning methods such as neural nets and support vector machines can achieve up to 80% of overall accuracy in globular proteins. In the case of the prediction of the final tertiary structure, the methods range from comparative modeling with resolved structures to ab initio modeling approaches. In comparative modeling, the search space is restricted by the assumption that the target protein adopts a structure close to the experimentally determined structure of another homologous protein.

In contrast to this, ab initio prediction is more challenging as it uses information from the amino acid sequence only. Typical approaches to ab initio prediction simplify the complexity of the interactions and the nature of the amino acid elements, e.g. through the use of lattice-based or low-resolution models such as the representation employed by the first stage of the Rosetta technique. Even with such simplified models, the energy landscapes in these problems present a multitude of local energy minima that are typically separated by high barriers. Given these challenges, recent research in this field has therefore focused on the development of improved sampling protocols for protein structure prediction, often using meta-heuristic techniques routed in the field of natural computation.

The aim of this workshop is to provide a forum for the exchange and communication of ideas, proposals and results related to the use of nature-inspired techniques in problems related to computational protein structure prediction. In tackling this important problem, nature-inspired techniques are currently being used in a variety of ways, but presentations related to this work are often distributed across a range of sessions / conferences / journals dependent on the particular sub-problem considered / algorithm used. It is hoped that this workshop will act as a meeting point for those authors and attendants of the PPSN conference who have a current or developing interest in this area.




Topic areas include (but are not restricted to):

  Use of natural and evolutionary computing algorithms in protein structure prediction (secondary and tertiary).

  Use of natural computing for determining protein classification and protein function.

  Hybrid combinations of algorithms applied to protein structure prediction or protein classification.

  Use of artificial life models like cellular automata or Lindenmayer systems.

  Use of simple and detailed lattice models and CASP (Critical Assessment of Techniques for Protein Structure Prediction) evaluation measures (RMSD, GDT, …).

  Multi-objective approaches.

  Modeling of temporal folding.

  Integration of visualization methods of protein structure within the process with natural computing alternatives.

  Use of surrogate models in order to reduce the computational time of current approaches.

  Parallel implementations (Threads, MPI, OpenMP, GPUs or FPGAs).



Format & Submission:

Authors interested in presenting their works should submit an extended abstract describing their work (max one A4 page). The workshop will run over a half day where the authors will have the opportunity for presenting and discussing their developments in the related topics.

A printed version of the abstracts will be provided to the workshop attendants.






Presentations and schedule:

Chair: José Santos

16:00-16:30 José Santos, Pablo Villot, Martin Diéguez, “Cellular automata for modeling protein folding in lattice models”.


16:30-17:00 Sune S. Nielsen, Wiktor Jurkowski, Grégoire Danoy, Juan Luis Jiménez Laredo, Reinhard Schneider, El-Ghazali Talbi and Pascal Bouvry, “Evolutionary multi objective optimisation with diversity as objective for the protein structure similarity problem”.


17:00-17:30 Shaun M. Kandathil, Simon C. Lovell and Julia Handl, “Low-resolution conformational exploration for Rosetta ab initio by bi-level optimisation of structural features”.


17:30-18:00 Amarda Shehu and Kenneth A. De Jong, “Memetic, multi-objective, off-lattice, and multiscale evolutionary algorithms for de novo and guided protein structure modelling”.






José Santos


University of A Coruña, Spain


Gregorio Toscano


Cinvestav, Information Technology Lab -Tamaulipas, México


Julia Handl


University of Manchester, UK