Computational peptide discovery with a genetic programming approach.

The development of peptides for therapeutic targets or biomarkers for disease diagnosis is a challenging task in protein engineering. Current approaches are tedious, often time-consuming and require complex laboratory data due to the vast search spaces that need to be considered. In silico methods can accelerate research and substantially reduce costs. Evolutionary algorithms are a promising approach for exploring large search spaces and can facilitate the discovery of new peptides. This study presents the development and use of a new variant of the genetic-programming-based POET algorithm, called POET Regex , where individuals are represented by a list of regular expressions. This algorithm was trained on a small curated dataset and employed to generate new peptides improving the sensitivity of peptides in magnetic resonance imaging with chemical exchange saturation transfer (CEST). The resulting model achieves a performance gain of 20% over the initial POET models and is able to predict a candidate peptide with a 58% performance increase compared to the gold-standard peptide. By combining the power of genetic programming with the flexibility of regular expressions, new peptide targets were identified that improve the sensitivity of detection by CEST. This approach provides a promising research direction for the efficient identification of peptides with therapeutic or diagnostic potential.

A Spatial Artificial Chemistry Implementation of a Gene Regulatory Network Aimed at Generating Protein Concentration Dynamics.

Gene regulatory networks are networks of interactions in organisms responsible for determining the production levels of proteins and peptides. Mathematical and computational models of gene regulatory networks have been proposed, some of them rather abstract and called artificial regulatory networks. In this contribution, a spatial model for gene regulatory networks is proposed that is biologically more realistic and incorporates an artificial chemistry to realize the interaction between regulatory proteins called the transcription factors and the regulatory sites of simulated genes. The result is a system that is quite robust while able to produce complex dynamics similar to what can be observed in nature. Here an analysis of the impact of the initial states of the system on the produced dynamics is performed, showing that such models are evolvable and can be directed toward producing desired protein dynamics.

Computational Peptide Discovery with a Genetic Programming Approach.

Background: The development of peptides for therapeutic targets or biomarkers for disease diagnosis is a challenging task in protein engineering. Current approaches are tedious, often time-consuming and require complex laboratory data due to the vast search space. In silico methods can accelerate research and substantially reduce costs. Evolutionary algorithms are a promising approach for exploring large search spaces and facilitating the discovery of new peptides. Results: This study presents the development and use of a variant of the initial POET algorithm, called POETRegex, which is based on genetic programming, where individuals are represented by a list of regular expressions. The program was trained on a small curated dataset and employed to predict new peptides that can improve the problem of sensitivity in detecting peptides through magnetic resonance imaging using chemical exchange saturation transfer (CEST). The resulting model achieves a performance gain of 20% over the initial POET variant and is able to predict a candidate peptide with a 58% performance increase compared to the gold-standard peptide. Conclusions: By combining the power of genetic programming with the flexibility of regular expressions, new potential peptide targets were identified to improve the sensitivity of detection by CEST. This approach provides a promising research direction for the efficient identification of peptides with therapeutic or diagnostic potential.

A Genetic Programming Approach to Engineering MRI Reporter Genes.

Here we develop a mechanism of protein optimization using a computational approach known as "genetic programming". We developed an algorithm called Protein Optimization Engineering Tool (POET). Starting from a small library of literature values, the use of this tool allowed us to develop proteins that produce four times more MRI contrast than what was previously state-of-the-art. Interestingly, many of the peptides produced using POET were dramatically different with respect to their sequence and chemical environment than existing CEST producing peptides, and challenge prior understandings of how those peptides function. While existing algorithms for protein engineering rely on divergent evolution, POET relies on convergent evolution and consequently allows discovery of peptides with completely different sequences that perform the same function with as good or even better efficiency. Thus, this novel approach can be expanded beyond developing imaging agents and can be used widely in protein engineering.

Long-Term Evolution Experiment with Genetic Programming.

We evolve floating point Sextic polynomial populations of genetic programming binary trees for up to a million generations. We observe continued innovation but this is limited by tree depth. We suggest that deep expressions are resilient to learning as they disperse information, impeding evolvability, and the adaptation of highly nested organisms, and we argue instead for open complexity. Programs with more than 2,000,000,000 instructions (depth 20,000) are created by crossover. To support unbounded long-term evolution experiments in genetic programming (GP), we use incremental fitness evaluation and both SIMD parallel AVX 512-bit instructions and 16 threads to yield performance equivalent to 1.1 trillion GP operations per second, 1.1 tera GPops, on an Intel Xeon Gold 6136 CPU 3.00GHz server.

From Dynamics to Novelty: An Agent-Based Model of the Economic System.

The modern economy is both a complex self-organizing system and an innovative, evolving one. Contemporary theory, however, treats it essentially as a static equilibrium system. Here we propose a formal framework to capture its complex, evolving nature. We develop an agent-based model of an economic system in which firms interact with each other and with consumers through market transactions. Production functions are represented by a pair of von Neumann technology matrices, and firms implement production plans taking into account current price levels for their inputs and output. Prices are determined by the relation between aggregate demand and supply. In the absence of exogenous perturbations the system fluctuates around its equilibrium state. New firms are introduced when profits are above normal, and are ultimately eliminated when losses persist. The varying number of firms represents a recurrent perturbation. The system thus exhibits dynamics at two levels: the dynamics of prices and output, and the dynamics of system size. The model aims to be realistic in its fundamental structure, but is kept simple in order to be computationally efficient. The ultimate aim is to use it as a platform for modeling the structural evolution of an economic system. Currently the model includes one form of structural evolution, the ability to generate new technologies and new products.

The effects of taxes on wealth inequality in Artificial Chemistry models of economic activity.

We consider a number of Artificial Chemistry models for economic activity and what consequences they have for the formation of economic inequality. We are particularly interested in what tax measures are effective in dampening economic inequality. By starting from well-known kinetic exchange models, we examine different scenarios for reducing the tendency of economic activity models to form unequal wealth distribution in equilibrium.

Neural Architecture Transfer.

Neural architecture search (NAS) has emerged as a promising avenue for automatically designing task-specific neural networks. Existing NAS approaches require one complete search for each deployment specification of hardware or objective. This is a computationally impractical endeavor given the potentially large number of application scenarios. In this paper, we propose Neural Architecture Transfer (NAT) to overcome this limitation. NAT is designed to efficiently generate task-specific custom models that are competitive under multiple conflicting objectives. To realize this goal we learn task-specific supernets from which specialized subnets can be sampled without any additional training. The key to our approach is an integrated online transfer learning and many-objective evolutionary search procedure. A pre-trained supernet is iteratively adapted while simultaneously searching for task-specific subnets. We demonstrate the efficacy of NAT on 11 benchmark image classification tasks ranging from large-scale multi-class to small-scale fine-grained datasets. In all cases, including ImageNet, NATNets improve upon the state-of-the-art under mobile settings ( ≤ 600M Multiply-Adds). Surprisingly, small-scale fine-grained datasets benefit the most from NAT. At the same time, the architecture search and transfer is orders of magnitude more efficient than existing NAS methods. Overall, experimental evaluation indicates that, across diverse image classification tasks and computational objectives, NAT is an appreciably more effective alternative to conventional transfer learning of fine-tuning weights of an existing network architecture learned on standard datasets. Code is available at https://github.com/human-analysis/neural-architecture-transfer.

Prediction of normalized signal strength on DNA sequencing microarrays by n-grams within a neural network model.

We have shown previously that a feed-forward, back propagation neural network model based on composite n-grams can predict normalized signal strengths of a microarray based DNA sequencing experiment. The microarray comprises a 4xN set of 25-base single-stranded DNA molecule ('oligos'), specific for each of the four possible bases (A, C, G, or T) for Adenine, Cytosine, Guanine and Thymine respectively at each of N positions in the experimental DNA. Strength of binding between reference oligos and experimental DNA varies according to base complementarity and the strongest signal in any quartet should `call the base` at that position. Variation in base composition of and (or) order within oligos can affect accuracy and (or) confidence of base calls. To evaluate the effect of order, we present oligos as n-gram neural input vectors of degree 3 and measure their performance. Microarray signal intensity data were divided into training, validation and testing sets. Regression values obtained were >99.80% overall with very low mean square errors that transform to high best validation performance values. Pattern recognition results showed high percentage confusion matrix values along the diagonal and receiver operating characteristic curves were clustered in the upper left corner, both indices of good predictive performance. Higher order n-grams are expected to produce even better predictions.

Artificial Gene Regulatory Networks-A Review.

In nature, gene regulatory networks are a key mediator between the information stored in the DNA of living organisms (their genotype) and the structural and behavioral expression this finds in their bodies, surviving in the world (their phenotype). They integrate environmental signals, steer development, buffer stochasticity, and allow evolution to proceed. In engineering, modeling and implementations of artificial gene regulatory networks have been an expanding field of research and development over the past few decades. This review discusses the concept of gene regulation, describes the current state of the art in gene regulatory networks, including modeling and simulation, and reviews their use in artificial evolutionary settings. We provide evidence for the benefits of this concept in natural and the engineering domains.

Use of a neural network to predict normalized signal strengths from a DNA-sequencing microarray.

A microarray DNA sequencing experiment for a molecule of N bases produces a 4xN data matrix, where for each of the N positions each quartet comprises the signal strength of binding of an experimental DNA to a reference oligonucleotide affixed to the microarray, for the four possible bases (A, C, G, or T). The strongest signal in each quartet should result from a perfect complementary match between experimental and reference DNA sequence, and therefore indicate the correct base call at that position. The linear series of calls should constitute the DNA sequence. Variation in the absolute and relative signal strengths, due to variable base composition and other factors over the N quartets, can interfere with the accuracy and (or) confidence of base calls in ways that are not fully understood. We used a feed-forward back-propagation neural network model to predict normalized signal intensities of a microarray-derived DNA sequence of N = 15,453 bases. The DNA sequence was encoded as n-gram neural input vectors, where n = 1, 2, and their composite. The data were divided into training, validation, and testing sets. Regression values were >99% overall, and improved with increased number of neurons in the hidden layer, and in the composition n-grams. We also noticed a very low mean square error overall which transforms to a high performance value.

Open-Ended Evolution: Perspectives from the OEE Workshop in York.

We describe the content and outcomes of the First Workshop on Open-Ended Evolution: Recent Progress and Future Milestones (OEE1), held during the ECAL 2015 conference at the University of York, UK, in July 2015. We briefly summarize the content of the workshop's talks, and identify the main themes that emerged from the open discussions. Two important conclusions from the discussions are: (1) the idea of pluralism about OEE-it seems clear that there is more than one interesting and important kind of OEE; and (2) the importance of distinguishing observable behavioral hallmarks of systems undergoing OEE from hypothesized underlying mechanisms that explain why a system exhibits those hallmarks. We summarize the different hallmarks and mechanisms discussed during the workshop, and list the specific systems that were highlighted with respect to particular hallmarks and mechanisms. We conclude by identifying some of the most important open research questions about OEE that are apparent in light of the discussions. The York workshop provides a foundation for a follow-up OEE2 workshop taking place at the ALIFE XV conference in Cancún, Mexico, in July 2016. Additional materials from the York workshop, including talk abstracts, presentation slides, and videos of each talk, are available at http://alife.org/ws/oee1 .

Defining and simulating open-ended novelty: requirements, guidelines, and challenges.

The open-endedness of a system is often defined as a continual production of novelty. Here we pin down this concept more fully by defining several types of novelty that a system may exhibit, classified as variation, innovation, and emergence. We then provide a meta-model for including levels of structure in a system's model. From there, we define an architecture suitable for building simulations of open-ended novelty-generating systems and discuss how previously proposed systems fit into this framework. We discuss the design principles applicable to those systems and close with some challenges for the community.

The effects of recombination on phenotypic exploration and robustness in evolution.

Recombination is a commonly used genetic operator in artificial and computational evolutionary systems. It has been empirically shown to be essential for evolutionary processes. However, little has been done to analyze the effects of recombination on quantitative genotypic and phenotypic properties. The majority of studies only consider mutation, mainly due to the more serious consequences of recombination in reorganizing entire genomes. Here we adopt methods from evolutionary biology to analyze a simple, yet representative, genetic programming method, linear genetic programming. We demonstrate that recombination has less disruptive effects on phenotype than mutation, that it accelerates novel phenotypic exploration, and that it particularly promotes robust phenotypes and evolves genotypic robustness and synergistic epistasis. Our results corroborate an explanation for the prevalence of recombination in complex living organisms, and helps elucidate a better understanding of the evolutionary mechanisms involved in the design of complex artificial evolutionary systems and intelligent algorithms.

Reducing the number of fitness evaluations in graph genetic programming using a canonical graph indexed database.

In this paper we describe the genetic programming system GGP operating on graphs and introduce the notion of graph isomorphisms to explain how they influence the dynamics of GP. It is shown empirically how fitness databases can improve the performance of GP and how mapping graphs to a canonical form can increase these improvements by saving considerable evaluation time.

Guidelines: From artificial evolution to computational evolution: a research agenda.

Computational scientists have developed algorithms inspired by natural evolution for at least 50 years. These algorithms solve optimization and design problems by building solutions that are 'more fit' relative to desired properties. However, the basic assumptions of this approach are outdated. We propose a research programme to develop a new field: computational evolution. This approach will produce algorithms that are based on current understanding of molecular and evolutionary biology and could solve previously unimaginable or intractable computational and biological problems.

Network topology and the evolution of dynamics in an artificial genetic regulatory network model created by whole genome duplication and divergence.

Topological measures of large-scale complex networks are applied to a specific artificial regulatory network model created through a whole genome duplication and divergence mechanism. This class of networks share topological features with natural transcriptional regulatory networks. Specifically, these networks display scale-free and small-world topology and possess subgraph distributions similar to those of natural networks. Thus, the topologies inherent in natural networks may be in part due to their method of creation rather than being exclusively shaped by subsequent evolution under selection. The evolvability of the dynamics of these networks is also examined by evolving networks in simulation to obtain three simple types of output dynamics. The networks obtained from this process show a wide variety of topologies and numbers of genes indicating that it is relatively easy to evolve these classes of dynamics in this model.

Dynamic subset selection based on a fitness case topology.

A large training set of fitness cases can critically slow down genetic programming, if no appropriate subset selection method is applied. Such a method allows an individual to be evaluated on a smaller subset of fitness cases. In this paper we suggest a new subset selection method that takes the problem structure into account, while being problem independent at the same time. In order to achieve this, information about the problem structure is acquired during evolutionary search by creating a topology (relationship) on the set of fitness cases. The topology is induced by individuals of the evolving population. This is done by increasing the strength of the relation between two fitness cases, if an individual of the population is able to solve both of them. Our new topology-based subset selection method chooses a subset, such that fitness cases in this subset are as distantly related as is possible with respect to the induced topology. We compare topology-based selection of fitness cases with dynamic subset selection and stochastic subset sampling on four different problems. On average, runs with topology-based selection show faster progress than the others.

Microarray-based in vitro evaluation of DNA oligomer libraries designed in silico.

We report on the microarray-based in vitro evaluation of two libraries of DNA oligonucleotide sequences, designed in silico for applications in supramolecular self-assembly, such as DNA computing and DNA-based nanosciences. In this first study which is devoted to the comparison of sequence motif properties theoretically predicted with their performance in real-life, the DNA-directed immobilization (DDI) of proteins was used as an example of DNA-based self-assembly. Since DDI technologies, DNA computing, and DNA nanoconstruction essentially depend on similar prereguisites, in particular, large and uniform hybridization efficiencies combined with low nonspecific cross-reactivity between individual sequences, we anticipate that the microarray approach demonstrated here will enable rapid evaluation of other DNA sequence libraries.