Sune Steinbjorn Nielsen

A Novel Multi-objectivisation Approach for Optimising the Protein Inverse Folding Problem

In biology, the subject of protein structure prediction is of continued interest, not only to chart the molecular map of the living cell, but also to design proteins of new functions. The Inverse Folding Problem (IFP) is in itself an important research problem, but also at the heart of most rational protein design approaches. In brief, the IFP consists in finding sequences that will fold into a given structure, rather than determining the structure for a given sequence - as in conventional structure prediction. In this work we present a Multi Objective Genetic Algorithm (MOGA) using the diversity-as-objective (DAO) variant of multi-objectivisation, to optimise secondary structure similarity and sequence diversity at the same time, hence pushing the search farther into wide-spread areas of the sequence solution-space. To control the high diversity generated by the DAO approach, we add a novel Quantile Constraint (QC) mechanism to discard an adjustable worst quantile of the population. This DAO-QC approach can efficiently emphasise exploitation rather than exploration to a selectable degree achieving a trade-off producing both better and more diverse sequences than the standard Genetic Algorithm (GA). To validate the final results, a subset of the best sequences was selected for tertiary structure prediction. The super-positioning with the original protein structure demonstrated that meaningful sequences are generated underlining the potential of this work.

Novel efficient asynchronous cooperative co-evolutionary multi-objective algorithms

This article introduces asynchronous implementations of selected synchronous cooperative co-evolutionary multi-objective evolutionary algorithms (CCMOEAs). The CCMOEAs chosen are based on the following state-of-the-art multi-objective evolutionary algorithms (MOEAs): Non-dominated Sorting Genetic Algorithm II (NSGA-II), Strength Pareto Evolutionary Algorithm 2 (SPEA2) and Multi-objective Cellular Genetic Algorithm (MOCell). The cooperative co-evolutionary variants presented in this article differ from the standard MOEAs architecture in that the population is split into islands, each of them optimizing only a sub-vector of the global solution vector, using the original multi-objective algorithm. Each island evaluates complete solutions through cooperation, i.e., using a subset of the other islands current partial solutions. We propose to study the performance of the asynchronous CCMOEAs with respect to their synchronous versions and base MOEAs on well kown test problems, i.e. ZDT and DTLZ. The obtained results are analyzed in terms of both the quality of the Pareto front approximations and computational speedups achieved on a multicore machine.

Cooperative Selection: Improving Tournament Selection via Altruism

This paper analyzes the dynamics of a new selection scheme based on altruistic cooperation between individuals. The scheme, which we refer to as cooperative selection, extends from tournament selection and imposes a stringent restriction on the mating chances of an individual during its lifespan: winning a tournament entails a depreciation of its fitness value. We show that altruism minimizes the loss of genetic diversity while increasing the selection frequency of the fittest individuals. An additional contribution of this paper is the formulation of a new combinatorial problem for maximizing the similarity of proteins based on their secondary structure. We conduct experiments on this problem in order to validate cooperative selection. The new selection scheme outperforms tournament selection for any setting of the parameters and is the best trade-off, maximizing genetic diversity and minimizing computational efforts.

Tackling the IFP Problem with the Preference-Based Genetic Algorithm

In molecular biology, the subject of protein structure prediction is of continued interest, not only to chart the molecular map of living cells, but also to design proteins with new functions. The Inverse Folding Problem (IFP) of finding sequences that fold into a defined structure is in itself an important research problem at the heart of rational protein design. In this work the Preference-Based Genetic Algorithm (PBGA) is employed to find many diversified solutions to the IFP. The PBGA algorithm incorporates a weighted sum model in order to combine fitness and diversity into a single objective function scoring a set of individuals as a whole. By adjusting the sum weights, a direct control of the fitness vs. diversity trade-off in the algorithm population is achieved by means of a selection scheme iteratively removing the least contributing individuals. Experimental results demonstrate the superior performance of the PBGA algorithm compared to other state-of-the-art algorithms both in terms of fitness and diversity.

NK Landscape Instances Mimicking the Protein Inverse Folding Problem Towards Future Benchmarks

This paper introduces two new nominal NK Landscape model instances designed to mimic the properties of one challenging optimisation problem from biology: the Inverse Folding Problem (IFP), here focusing on a simpler secondary structure version. Through landscape analysis tests, numerous problem properties are identified and used to parameterise and validate model instances in terms of epistatic links, adaptive- and random walk characteristics. Then the performance of different Genetic Algorithms (GAs) is compared on both the new NK Models and the original IFP, in terms of population diversity, solution quality and convergence characteristics. It is demonstrated that very similar properties are captured in all presented tests with a significantly faster evaluation time compared to the real IFP. The future purpose of such a model is to provide a generic benchmark for algorithms targeting protein sequence optimisation, specifically in protein design. It may also provide the foundation for more in-depth studies of the size, shape and characteristics of the solution space of good solutions to the IFP.

Vehicular mobility model optimization using cooperative coevolutionary genetic algorithms

A key factor for accurate vehicular ad hoc networks (VANET) simulation is the quality of its underlying mobility model. VehILux is a recent vehicular mobility model that generates traces using traffic volume counts and real-world map data. This model uses probabilistic attraction points which values require optimization to provide realistic traces. Previous sensitivity analysis and application of genetic algorithms (GAs) on the Luxembourg problem instance have outlined this models limitations. In this article, we first propose an extension of the model using a higher number of auto-generated attraction points. Then its decomposition on the Luxembourg instance using geographical information is proposed as a way to break epistatic links and hence make its optimization using cooperative coevolutionary genetic algorithms (CCGAs) more efficient. Experimental results demonstrate the significant realism increase brought by both the VehILux model enhancements and the CCGA compared to the generational and cellular GAs.