The field of microbiome research has experienced exponential growth in recent years, driven by advance ments in next-generation sequencing technologies. A particular topic of interest in this field lies in studying the influence of the microbiome on the host organism's state, with the aim of boosting the potential use of microorganisms as indicators of phenotypic conditions. However, microorganism selection is a hard-to-tackle problem involving multiple optimality criteria. This work introduces MADMISM, a Multi-objective Approach based on Dominance for MIcroorganism Selection on Microbiomes. Based on the Non-dominated Sorting Genetic Algorithm III, the proposed approach includes maximum-entropy classification with stochastic gradient descent learning and domain-specific operators through all evolutionary stages. Hence, MADMISM combines the search power of evolutionary algorithms with the enhanced guidance provided by the domain-specific operators. The evaluation of results on eight real microbiome datasets showcases MADMISM's superiority in providing solutions with a minimal set of Operational Taxonomic Units (OTUs) while achieving high classification performance, with statistically significant improvements over state-of-the-art methods in most scenarios. Specifically, MADMISM accomplishes an average OTU reduction of 46.35 % and a classification improvement of 12.05 % compared to state-of-the-art methods, while also showcasing remarkable scalability. Furthermore, MADMISM exhibited supe rior multi-objective performance with an average hypervolume of 90.06 %, and presented statistically significant improvements over well-established multi-objective algorithms. Therefore, MADMISM represents an effective ap proach for identifying microbial markers. This is evidenced by the quality of its results and the biological relevance of the selected microorganisms. Materials for this work can be found at https://github.com/frodbej/madmism.

Gene networks have gained considerable relevance in cancer research, enabling the representation of complex biological relationships that provide insights into the mechanisms driving tumor development and progression. The increasing availability of biological data facilitates the construction of clinically relevant gene networks by integrating multiple information sources. Specifically, we consider mutation data, patient survival data, and protein-protein interaction data to identify networks whose genes are recurrently mutated, significantly involved in patient survival, and functionally associated. To this end, we apply multi-objective optimization to simultaneously maximize survival impact, functional association, and mutation coverage. Herein, we introduce MOTEA-GENSU (Multi-Objective Two-archive Evolutionary Algorithm to discover GEne Networks involved in SUrvival), a novel method that employs two collaborative archives and intelligent evolutionary operators to guide the generation of high-quality gene networks. Evaluation across 27 real biological scenarios covering diverse cancer types shows that MOTEA-GENSU outperforms existing methods, achieving superior results in 92.6% of comparisons, with improvements of up to 315.8% over the best-performing competing approach, and consistently surpassing all state-of-the-art methods on average within each evaluated dataset. Biological analysis of the identified networks validates their functional coherence and significant impact on cancer patient survival, revealing clinically relevant networks composed of genes with demonstrated prognostic value.

At present, many decisions, even decisions that affect people's lives, are increasingly made by decision-making software. For this reason, it is really important to avoid decision-making software being biased, that is, it needs to assure fairness. As decision-making software is generally based on classification models, its bias can be mitigated in three different stages: pre-processing, in-processing, and post-processing. However, despite the importance of fairness on these models, there are very few proposals for the post-processing stage that are able to mitigate bias without reducing the original model's accuracy. Therefore, this problem should be addressed as a multi-objective problem, optimizing at the same time both fairness and accuracy. Taking this into account, we propose a Multi-Objective Swarm Intelligence approach for BIas Mitigation (MOSIBIM), which combines dominance-based multi-objective optimization (with techniques such as Pareto fronts, niching, and reference points), population-based evolutionary computation, swarm intelligence, and convergence-stagnation differentiation. In order to analyze the improvements that this proposal produces over other approaches in the literature, its results on fairness and accuracy have been compared with the results obtained by other five approaches, optimizing different classification models and in six distinct real-life scenarios which have various types of bias. The results of MOSIBIM show great improvements on fairness in comparison with the other approaches, as it reaches improvement percentages of 91.7% and, on average, the results are always in the range of 66.6% to 76.6% of improvement. Furthermore, the proposed approach has been able to improve the original model's accuracy in all the studied cases.

The use of keywords is increasingly being applied across diverse domains, including the movie industry, whose main platforms are adopting advanced natural language processing techniques. Algorithms for automatic extraction of keywords can provide relevant information in this domain. The most novel approaches covering several categories (statistics, graphs, word embedding, and hybrid) have been considered in a model study framework. They have been implemented, applied, and evaluated with standard datasets. In addition, a movie dataset with gold standard keywords, based on textual metadata from synopses and reviews, has been specifically developed for this scope. Keyword extraction models have been evaluated in terms of F-score and computation time. Furthermore, content analysis, both quantitative and qualitative, of the extracted keywords in the movie context has been performed. Results show a great variability in model performance and computation time among the different models. Qualitative results, in addition to F-score and computation time, demonstrate that keyword extraction works better with synopses than with reviews. The quantitative content analysis revealed that EmbedRank effectively reduces redundancy and limits the use of proper nouns, leading to high-quality keywords.

In the pursuit of sustainable computational approaches, research on parallel computing has been progressively orienting efforts towards the definition of energy-aware methodologies to solve real-world problems. A commonly adopted approach in this context lies in the adjustment of operating frequencies to seek power-time tradeoffs. However, the characteristics of the applications impose restrictions on the benefits that can be achieved through this technique, thus demanding the design of accurate algorithmic strategies to enhance energy efficiency. This work explores the combination of frequency scaling and multi-threaded vector-based algorithms to improve energy consumption in an important problem from the evolutionary biology domain: phylogenetic inference. Different technologies, such as AVX512, AVX2, and GPU vector types, are investigated to define refined parallel designs based on OpenMP, for CPUs, and CUDA, for GPUs. Experimental results on five biological datasets give account of energy improvements of up to 20 × when tackling complex problem instances on CPUs, while also observing significant benefits on different GPU processing scenarios. Consequently, the proposed strategies effectively facilitate the definition of energy-aware execution environments for phylogenetics, improving other literature approaches in terms of time and energy consumption.

Motivation:Phage therapy is emerging as a promising alternative to antibiotics in biomedical research, highlighting the growing need for computational tools to rationally design effective phage cocktails. However, its clinical potential is often compromised by the evolution of heritable bacterial resistance, which is frequently exacerbated by repeated phage exposure. This can lead to broad-spectrum cross-resistance and reduced long-term efficacy. Existing approaches typically rely on host range matrices but often overlook viral interference and the complex, non-binary nature of virus-host interactions. Results:We present SocialViruses, a tool for designing optimized phage cocktails selecting up to twelve viruses and using two alternative algorithms. SocialViruses integrates quantitative host range infection and virus-virus interaction matrices to guide cocktail design. It produces a detailed report with key quality metrics and allows users to define multiple cocktails while minimizing viral interference and managing co-infection redundancy. Availability and implementation:SocialViruses is freely available as a Cytoscape application and can be downloaded from: https://apps.cytoscape.org/apps/SocialViruses.

The explosive growth of uncategorized web documents requires effective methods for their organization. Clustering techniques address this challenge by automatically grouping similar documents, making large amounts of unstructured data more manageable. In recent times, multi-objective optimization approaches have become an effective way to solve the document clustering problem. However, in the scientific literature, there is a lack of studies that address this task by using swarm-intelligence algorithms from this optimization viewpoint. For this reason, a multi-objective swarm-intelligence algorithm for document clustering (MOSIDOC) has been designed, developed, and applied in this work. MOSIDOC is based on the idea of combining swarm-intelligence mechanisms from artificial bee colony with problem-aware operators to attain an accurate processing of the search space. The criteria of compactness, separation, and Davies-Bouldin index have been formulated as the objective functions to be optimized. The experimentation has been carried out on ODP-239, one of the most widely-used datasets for testing document clustering methods. To comprehensively evaluate the proposed approach, the evaluation metrics of cluster-level F1 measure, Adjusted Rand Index, and Normalized Mutual Information have been applied. The obtained results denote that MOSIDOC leads to average percentage improvements up to 86.81%, in comparison to other fifteen competing methods.

Identifying cancer driver pathways is essential in cancer research. This problem is well-suited to multi-objective optimization, as cancer driver pathways consist of frequently mutated genes that cover a significant proportion of patients and exhibit high correlation due to their coordinated functionality. Therefore, this problem involves two competing objectives to maximize: patient coverage and gene network correlation. To optimize these objectives simultaneously, we propose a Multi-objective Artificial Bee Colony approach for Identifying cancer driver Pathways (MABCIP). Our approach leverages the strengths of the Artificial Bee Colony algorithm in combining exploitation (via employed and onlooker bees) and exploration (through scout bees). MABCIP incorporates problem-aware initialization and mutation operators which substantially enhance its performance. Specifically, MABCIP achieves noteworthy improvements in hypervolume and attains 100% set coverage inmost cases over the version with standard operators. We thoroughly evaluate MABCIP against state-ofthe-art methods using real cancer datasets comprehending 45 experimental scenarios. MABCIP consistently outperforms existing methods, attaining the best results in 44 out of 45 experimental scenarios and showing remarkable performance in complex instances. Notably, MABCIP obtains results up to 7.1 times better than those of the best-performing state-of-the-art method. Moreover, MABCIP identifies pathways involved in the development of targeted therapies, highlighting its potential for identifying clinically relevant mechanisms.

BACKGROUND AND OBJECTIVE:Single-cell RNA-sequencing (scRNA-seq) has revolutionized transcriptomic studies by providing detailed insights into gene expression profiles at the single-cell level. This technology allows researchers to capture expression patterns of thousands of genes across hundreds or thousands of individual cells. Clustering is a crucial step in the analysis of scRNA-seq data, since it enables the identification of distinct cell populations based on their transcriptomic profiles and serves as a foundation for downstream analysis. Given that clustering scRNA-seq data is a challenging task that involves different conflicting objectives, our goal is to tackle it from a multi-objective optimization perspective. METHODS:This study proposes a Reference Vector-guided Evolutionary Algorithm for Cluster Analysis of Single-cell Transcriptomes (RVEA-CAST) to address the clustering task as a multi-objective optimization problem. Our approach considers three objectives to optimize: clustering deviation, clustering compactness, and the Davies-Bouldin index. The algorithmic design of RVEA-CAST incorporates three problem-aware mutation operators specifically designed to improve each objective, which are orchestrated under a multi-objective search engine based on the use of reference vectors. RESULTS:RVEA-CAST is evaluated on ten real scRNA-seq datasets using standard clustering evaluation metrics, such as Normalized Mutual Information (NMI) and Adjusted Rand Index (ARI). The attained results reveal the improved performance and robustness of the proposed approach compared to other previously proposed methods. Specifically, statistically significant improvements of up to 66.7% and 261.5% were achieved for NMI and ARI, respectively. Furthermore, the analysis of differentially expressed genes in the predicted and real clusters showcased greater agreement of our solutions with actual cell populations, underscoring the biological relevance of our approach. CONCLUSIONS:The results highlight that RVEA-CAST is an effective and versatile approach for clustering scRNA-seq data, outperforming existing methods across diverse biological scenarios in both widely used clustering evaluation metrics and biological relevance.

Background and ObjectiveThe protein-protein interaction (PPI) network alignment has proven to be an efficient technique in the diagnosis and prevention of certain diseases. However, the difficulty in maximizing, at the same time, the two qualities that measure the goodness of alignments (topological and biological quality) has led aligners to produce very different alignments. Thus making a comparative study among alignments of such different qualities a big challenge. Multi-objective optimization is a computer method, which is very powerful in this kind of contexts because both conflicting qualities are considered together. Analysing the alignments of each PPI network aligner with multi-objective methodologies allows you to visualize a bigger picture of the alignments and their qualities, obtaining very interesting conclusions. This paper proposes a comprehensive PPI network aligner study in the multi-objective domain.MethodsAlignments from each aligner and all aligners together were studied and compared to each other via Pareto dominance methodologies. The best alignments produced by each aligner and all aligners together for five different alignment scenarios were displayed in Pareto front graphs. Later, the aligners were ranked according to the topological, biological, and combined quality of their alignments. Finally, the aligners were also ranked based on their average runtimes.ResultsRegarding aligners constructing the best overall alignments, we found that SAlign, BEAMS, SANA, and HubAlign are the best options. Additionally, the alignments of best topological quality are produced by: SANA, SAlign, and HubAlign aligners. On the contrary, the aligners returning the alignments of best biological quality are: BEAMS, TAME, and WAVE. However, if there are time constraints, it is recommended to select SAlign to obtain high topological quality alignments and PISwap or SAlign aligners for high biological quality alignments.ConclusionsThe use of the SANA aligner is recommended for obtaining the best alignments of topological quality, BEAMS for alignments of the best biological quality, and SAlign for alignments of the best combined topological and biological quality. Simultaneously, SANA and BEAMS have above-average runtimes. Therefore, it is suggested, if necessary due to time restrictions, to choose other, faster aligners like SAlign or PISwap whose alignments are also of high quality.

Currently, automatic multi-document summarization is an interesting subject in numerous fields of study. As a part of it, query-focused summarization is becoming increasingly important in recent times. These methods can automatically produce a summary based on a query given by the user, including the most relevant information from the query at the same time as the redundancy among sentences is reduced. This can be achieved by developing and applying a multi-objective optimization approach. In this paper, an Indicator-based Multi-Objective Variable Neighborhood Search (IMOVNS) algorithm has been designed, implemented, and tested for the query-focused extractive multi-document summarization problem. Experiments have been carried out with datasets from Text Analysis Conference (TAC). The results were evaluated using the Recall-Oriented Understudy for Gisting Evaluation (ROUGE) metrics. IMOVNS has greatly improved the results presented in the scientific literature, providing improvement percentages in ROUGE metric reaching up to 69.24% in ROUGE-1, up to 57.70% in ROUGE-2, and up to 77.37% in ROUGE-SU4 scores. Hence, the proposed IMOVNS offers a promising solution to the query-focused summarization problem, thus highlighting its efficacy and potential for enhancing automatic summarization techniques.

The amplification of the 16S ribosomal RNA gene through the polymerase chain reaction (PCR) is the main approach to profile bacterial communities. This gene is a widely used marker with a composition that allows the identification of microorganisms at the genus or species levels. The correct performance of a PCR assay depends on the properties of the chosen set of primers that match with the target 16S sequences, allowing the amplification by a DNA polymerase. For this reason, optimizing the design of primers attending to such multiple properties is crucial to ensure the specificity and robustness of this process. However, only one multiobjective proposal exists that addresses the optimization of primer design targeted to the 16S gene amplification. Herein, we propose a novel approach for multiobjective primer optimization based on a mutation-based genetic algorithm with three new problem-aware mutation operators, each aimed at covering one of the three objectives that compose the problem (efficiency, coverage, and variability). The proposed algorithm has been tested on 5 real datasets of bacterial 16S gene sequences. The results have been evaluated using 4 quality metrics, showing that our approach achieves statistically significant improvements with regard to the reference multiobjective approach in the field.

Morality is a topic that people are increasingly concerned about. Morality is observed and measured during public acts or when developing and consuming products, such as movies. The Moral Foundations Theory (MFT) was developed to rigorously perform these measurements with the support of the Moral Foundations Dictionary (MFD). In this paper, a Word Embedding-based Moral Foundation Assignment (WEMFA) approach has been designed, implemented, and applied to the movie domain for multiple assignment of moral foundations. WEMFA may use any dictionary, and it has been applied to a movie collection generated from movie synopses. A comparison between WEMFA and MoralStrength, the only approach found in the scientific literature, has been carried out. The proposed approach provided a percentage improvement of 41.7% with respect to the best version of MoralStrength, which uses an extension of the original MFD almost 10 times larger in number of terms. In addition, an extension of the original MFD (MFD24) has been built by adding 14 new moral foundations to the 10 original ones, enriching the moral context. WEMFA provided a mean accuracy of 78% with MFD24 despite the increment of the number of moral foundations. Besides, new extended dictionaries or even totally different ones can be used with WEMFA, since it does not need any training.

In population medical genetics, the study of autosomal recessive disorders in highly endogamous populations is a major topic where calculating the inbreeding and relationship coefficients on mating networks is crucial. However, a challenge arises when dealing with large and complex mating networks, making their traversal difficult during the calculation process. For this calculation, we propose using Iterative Level-0 (IL0) as a new and faster algorithm that traverses mating networks more efficiently. The purpose of this work is to explain in detail the IL0 algorithm and prove its superiority by comparing it with two algorithms based on the best-known algorithms in the area: Depth First Search (DFS) and Breadth First Search (BFS). A Cytoscape application has been developed to calculate the inbreeding and relationship coefficients of individuals composing any mating network. In this application, the IL0 proposal together with DFS-based and BFS-based algorithms have been implemented. Any user can access this freely available Cytoscape application (https://apps.cytoscape.org/apps/inbreeding) that allows the comparison between the IL0 proposal and the best-known algorithms (based on DFS and BFS). In addition, a diverse set of mating networks has been collected in terms of complexity (number of edges) and species (humans, primates, and dogs) for the experiments. The runtime obtained by the IL0, DFS-based, and BFS-based algorithms when calculating the inbreeding and relationship coefficients proved the improvement of IL0. In fact, a speedup study reflected that the IL0 algorithm is 7.60 to 127.50 times faster than DFS-based and BFS-based algorithms. Moreover, a scalability study found that the growth of the IL0 runtime has a linear dependence on the number of edges of the mating network, while the DFS-based and BFS-based runtimes have a quadratic dependence. Therefore, the IL0 algorithm can solve the problem of calculating the inbreeding and relationship coefficients many times faster (up to 127.50) than the two algorithms based on the famous DFS and BFS. Furthermore, our results demonstrate that IL0 scales much better as the complexity of mating networks increases.

Streaming services are increasingly leveraging Artificial Intelligence (AI) technologies for improved content cataloging, user experiences in content discovery, and personalization. A significant challenge in this domain is the automated assignment of microgenres to movies. This study introduces and evaluates approaches based on clustering, topic modeling, and word embedding to address this task. The evaluation employs a preprocessed dataset containing movie-related data—title tags, synopses, genres, and reviews—alongside a predefined microgenre list. Comparisons of three activation functions (binary step, ramp, and sigmoid) gauge their effectiveness in augmenting microgenre tags. Results demonstrate the superiority of the word embedding approach over clustering and topic modeling in terms of mean accuracy. Even more, the word embedding approach stands as the sole fully automated solution. Analysis indicates that incorporating review-based tags introduces noise and undermines accuracy. Besides, the word embedding approach yields optimal outcomes using the sigmoid function, effectively doubling assigned tags while maintaining matching quality. This sheds light on the potential of word embedding methods within the movie domain.

Currently, the explosive growth of the information available on the Internet makes automatic text summarization systems increasingly important. A particularly relevant challenge is the update summarization task. Update summarization differs from traditional summarization in its dynamic nature. While traditional summarization is static, that is, the document collections about a specific topic remain unchanged, update summarization addresses dynamic document collections based on a specific topic. Therefore, update summarization consists of summarizing the new document collection under the assumption that the user has already read a previous summarization and only the new information is interesting. The multiobjective number-one-selection genetic algorithm (MONOGA) has been designed and implemented to address this problem. The proposed algorithm produces a summary that is relevant to the user’s given query, and it also contains updates information. Experiments were conducted on Text Analysis Conference (TAC) datasets, and Recall-Oriented Understudy for Gisting Evaluation (ROUGE) metrics were considered to assess the model performance. The results obtained by the proposed approach outperform those from the existing approaches in the scientific literature, obtaining average percentage improvements between 12.74% and 55.03% in the ROUGE scores.

The number of investigations attempting to align protein–protein interaction (PPI) networks has increased with the growth of studies focused on collecting PPI data. These works aim to identify conserved areas between species that are difficult to differentiate due to speciation. However, there is no standard approach to align PPI networks, and global aligners encounter difficulties in constructing alignments with high biological and structural quality. To address this issue, we propose an innovative ensemble technique that combines the strengths of aligners in the PPI network alignment field while avoiding their weaknesses. This approach reduces the spread of dispersion in so different individual global aligners and contributes to achieving a global standard that produces alignments of higher quality. This is possible thanks to the two branches composing our ensemble that aim to improve alignments in terms of biological or structural quality. In addition to a new heuristic replacing the second-level aligner in the biological quality-focused branch. Our approach achieves alignments of higher quality, as demonstrated through experiments with 10 different scenarios involving real data from 5 species. Our solutions outperform other individual aligners and ensemble techniques, like bagging, in terms of biological and structural quality. Moreover, the time required to perform the ensemble is minimal compared to that of individual aligners.

Peer assessment has traditionally represented a key tool to enhance active learning and critical thinking. However, the success of this approach is governed by different factors, which have been accentuated in recent years. The implementation of peer assessment is consequently a challenging task in the current context. This work investigates peer assessment strategies in seven Computer Engineering courses. Students’ performance and assessment accuracy are analyzed throughout five academic years, covering the transition from offline to online methodologies according to the evolution of educational environments. More specifically, peer and lecturer’s grades are examined to identify correlations or deviations in two execution phases. The first phase involves the analysis of in-class offline peer assessments during four academic years, integrated as part of continuous assessment tasks. The second phase deals with the evaluation of online peer assessments in 2021/2022, considering different platforms to manage submissions and reviews. The offline experience denotes statistical correlations between the grades assigned by the peers and the lecturer, while also revealing patterns that affected the performance of students. In addition, the switch to online methodologies does not significantly affect the assessments in courses that adopted peer strategies in the past. Finally, comparable results are obtained under single-blind and double-blind models after careful training.

The integration of multiple genes to maximize protein expression levels represents an important challenge in synthetic biology. This task relies on the definition of multiple protein-coding sequences, which must be as different as possible to avoid information loss. Proteins can be encoded in different ways, using synonymous codons that translate into the same amino acid. Some codons are better suited to the host than others, thus being preferable the use of the most fitting ones. However, adopting only the most highly adapted codons would lead to very similar coding sequences. An additional criterion is given by the fact that the designed sequences must contain a suitable guanine–cytosine (GC) ratio in accordance with the characteristics of the host organism. Therefore, this biological task requires the simultaneous optimization of several, conflicting objectives. This work proposes a novel multi-objective approach for protein encoding, which tackles the problem according to a new formulation based on three objective functions: codon adaptation index, Hamming distance between sequences, and GC content. Our work extends the recent Butterfly Optimization Algorithm to multi-objective contexts, integrating problem-specific operators to boost solution quality by covering the different aspects required for accurate protein encoding. Two key structures, a taboo list and a best solution list, are defined to conduct improved searches attending to the potential improvements that each solution in the population can promote. Experiments conducted on nine real-world proteins reveal the attainment of relevant solutions from different evaluation perspectives, showing significant improvements over other single and multi-objective methods from the literature.

The RNA inverse folding problem involves discovering a nucleotide sequence that folds into a desired target structure. Although numerous computational methods have been proposed over the years to tackle the problem, none have successfully solved the complete Eterna100 set. The Eterna100 set is widely recognized as a benchmark in this field. Therefore, there is still ample room for improvement in this area. This paper aims to address this challenge by introducing eM2dRNAs, an enhanced version of our previous approach called m2dRNAs, which is a multiobjective metaheuristic to design RNA sequences. By introducing eM2dRNAs, we aim to make significant advancements in RNA inverse folding. Our approach starts with the recursive decomposition of the target structure, simplifying the problem to be solved. We conducted a comparative study of our method against several published methods using the Eterna100 benchmark. The results showed that our proposal performs significantly better than the other methods across almost all metrics and categories considered, thus achieving our objective of improving the ability to solve the RNA inverse folding problem.