Cutting-edge genomic science is opening entirely new possibilities for studying DNA, especially in contexts where understanding biological reserves and hidden genetic factors is critical for evaluating soldiers' performance under pressure. In this work, we introduce a hybrid computational strategy that combines particle swarm optimization (PSO) with a genetic algorithm (GA) to identify previously undetectable genes within the genome. The proposed approach is specifically designed to dissect structurally complex regions, such as centromeres and duplicated segments, that have long remained beyond the reach of standard sequencing tools. In the pipeline, PSO performs a broad sweep across massive genomic datasets to highlight areas potentially associated with stress tolerance or physical strength. At the same time, the GA stage narrows the search by modeling regulatory and evolutionary constraints. Experiments on synthetic genomic profiles simulating the genomes of military personnel showed that the method accurately predicts gene variants associated with reactions in extreme environments. When benchmarked against conventional analytical techniques, the hybrid solution demonstrated markedly higher precision and faster processing, indicating strong potential for integration into personalized genomic assessments for defense applications.
Within contemporary genomics, where analytical techniques for studying DNA are rapidly evolving, the
search for gene candidates that may be masked within the genome has become increasingly significant,
particularly for applications involving military personnel. Individuals in active service experience
exceptional psychological and physical loads, yet conventional sequencing and annotation pipelines
often overlook highly repetitive or structurally complex genomic regions. It is precisely in these zones
that factors influencing stress tolerance, accelerated recovery, or heightened cognitive performance
may reside [
The latest breakthroughs in achieving complete human genome assemblies, which finally resolve long-missing fragments such as centromeres and segmental duplications, together comprising nearly 8% of the genome, enable far more detailed exploration of genetic elements that could shape a soldier’s physiological readiness. Moreover, research conducted in defense-oriented laboratories has already pinpointed uncommon genetic variants, including those afecting serotonin regulation, that correlate with superior adaptation to intense training conditions and may serve as predictors of performance in elite military units.
To overcome these limitations, we introduce a combined computational framework that integrates
particle swarm optimization (PSO) with a genetic algorithm (GA), designed specifically for the analysis
of genomic profiles in military contexts. PSO modeled after the coordinated movement of animal groups
allows rapid traversal of vast genomic search spaces and is highly efective at spotting promising gene
signals within structurally challenging areas where classical analytic tools often fail. In turn, the GA
component relies on evolutionary logic, applying iterative selection, crossover, and mutation to refine
candidate solutions, prevent convergence to suboptimal points, and account for functional relationships
among genes influencing traits such as stress adaptation or physical stamina [
By merging these two strategies, the system becomes a robust instrument for military-focused genomics: PSO conducts the broad, high-speed sweep of the genomic landscape to flag markers associated with operational performance, while GA delivers deeper fine-tuning of these findings, incorporating empirical evidence from real sequencing datasets, including those obtained from elite forces trainees.
This methodology becomes even more critical with the emergence of fully assembled human genome datasets, which have expanded the possibilities for military-related genomic studies. High-accuracy long-read technologies now make it feasible to examine genomic territories that were previously out of reach, enabling the discovery of genetic factors tied to psychological stability and physiological adaptation. Completing these repetitive and structurally dense regions not only extends the total amount of analyzable DNA but also uncovers variants that may help mitigate risks of post-traumatic stress disorder or refine training strategies for diferent categories of soldiers.
The proposed hybrid system is applicable not only to pinpoint protein-coding genes but also to uncover
regulatory sequences whose activity shifts under dangerous or high-pressure conditions [
When designing a combined optimization framework that fuses particle swarm optimization with a
genetic algorithm for genomic analysis, especially within military-oriented genetics, it is essential
to build on existing research. Numerous earlier works demonstrate how PSO and GA can be jointly
employed to solve high-dimensional, non-linear optimization tasks: PSO performs fast, wide-range
probing of the search landscape, while GA refines potential solutions and prevents premature
convergence by applying crossover and mutation operations [
Breakthroughs in whole-genome sequencing, most notably the Telomere-to-Telomere (T2T) initiative, have underscored the importance of properly examining highly repetitive genomic segments. Similar regions, which constitute a large share of human DNA, remained poorly characterized for decades because earlier technologies could not resolve them. Research enabled by ultra-long reads and refined assembly techniques now makes it possible to close long-standing gaps in the genomic landscape, thereby opening the door to detecting gene variants that were previously hidden.
In military genomics, the same methodological advances can be repurposed to identify markers
associated with psychological resilience, physiological stamina, and other traits crucial to assessing
service members’ operational readiness. A growing body of research also investigates the application
of mixed computational strategies to manage massive bioinformatics datasets, including statistical
techniques for filtering genomic variants and refining the alignment of long-read sequences. These
solutions significantly improve assembly fidelity and could be incorporated into the proposed framework
to boost the precision of detecting genes embedded in structurally challenging regions [
Taken together, these prior investigations form both the conceptual and applied basis for the hybrid method proposed in this study. They demonstrate the value of integrating swarm-based search strategies with evolutionary optimization for intricate genomic datasets and highlight why such combined frameworks are essential for advancing military-focused genetic research. This body of work ultimately acts as a launch point for refining and extending our own algorithmic approach.
To deploy the hybrid framework integrating particle swarm optimization with a genetic algorithm to
detect concealed genes in military genomes, a structured, multi-phase procedure was followed. The
ifrst stage involved preprocessing the genomic sequences to eliminate noise and standardize data, with
special attention to repetitive regions such as centromeres. Subsequently, the algorithm was executed on
datasets modeled after soldiers’ genomic profiles [
Table 1 outlines the key stages of preparation and optimization employed in the proposed framework.
After completing the PSO stage, the candidate regions were forwarded to the genetic algorithm for
more in-depth and fine-grained optimization. At this point, the GA not only refined the solutions
previously identified by the swarm but also reassessed their internal structure to ensure that the most
informative and stable patterns were preserved. The algorithm employed an elitist strategy, which
guaranteed that the top-performing candidates were retained across generations, preventing the loss of
high-quality solutions during the evolutionary process. In parallel, crossover and mutation operators
were applied to introduce controlled variability, enabling the exploration of alternative configurations
and the discovery of new potential solutions with higher fitness. This cycle was iterated multiple
times to ensure convergence and stability of the results. Validation was conducted by benchmarking
against established gene annotations, allowing evaluation of the method’s reliability under practical
conditions [
To strengthen the analysis, an additional validation stage was introduced to simulate high-stress scenarios akin to combat environments. This approach allowed assessment of the algorithm’s performance on datasets where regulatory element activity could change under pressure. The outcomes were then consolidated into a comprehensive model for forecasting relevant genetic markers. The stages of this validation process are detailed in Table 2.
Ultimately, the processed data were examined to pinpoint genes associated with stress resilience and
adaptive capacity. The approach demonstrated strong performance in managing structurally complex
genomic regions [
4.1. Schematic model of the algorithm
The conceptual framework of the proposed hybrid algorithm, designed to uncover concealed genes
within military-grade genomic profiles, is predicated on a rigorous analysis of synthetic yet biologically
representative datasets. Because actual genomic data from active service members is highly sensitive
and restricted due to privacy and security concerns, we utilized test datasets designed to replicate
the stochastic nature and structural complexity of authentic human genomes. The analytic process
commences with the ingestion of high-dimensional input vectors, which include detailed attributes
of specific genomic segments. These attributes encompass the region type (e.g., centromeres, which
are traditionally dificult to sequence), the precise sequence length in base pairs (bp), and preliminary
potential scores derived from initial raw reads [
These inputs are not merely random variables; they are derived from simulated samples containing specific variants known to correlate with phenotypic traits essential for defense applications, such as stress resilience, metabolic eficiency, and rapid recovery from physical trauma. This simulation approach ensures that the algorithm is evaluated on data that closely approximates the noise levels and structural ambiguities found in real-world Next-Generation Sequencing (NGS) measurements. Table 3 delineates the baseline characteristics of the test data used for this analysis, highlighting the diversity of genomic structures subjected to the optimization pipeline.
Upon the successful loading of these test datasets, the algorithm initiates its primary exploration phase. In this stage, Particle Swarm Optimization (PSO) performs a global search across the defined genomic landscape. The PSO component treats the identification of gene candidates as a multi-objective optimization problem, conducting a broad sweep to flag regions that exhibit statistical anomalies indicative of functional coding sequences. This swarm-based approach is particularly efective at navigating local optima found in repetitive regions, such as sample S003.
Once the PSO layer identifies high-probability regions, these candidates are passed to the Genetic
Algorithm (GA) for granular refinement. The GA mimics biological evolution by applying operators
such as elitist selection, crossover, and mutation to the candidate solutions. This secondary stage is
crucial for modeling regulatory constraints and evolutionary conservation, ensuring that the detected
patterns are biologically plausible rather than mathematical artifacts [
In the second stage of the pipeline, the hybrid framework fundamentally transforms the initial metadata. It updates the candidate potential scores based on the convergence of the GA, generates specific functional gene predictions, and calculates a confidence metric expressed as a percentage. This multi-layered process demonstrates how the algorithm systematically adjusts starting values—often noisy or uncertain in raw data—to enhance the precision of detecting genes associated with critical military traits. For instance, the algorithm is able to re-evaluate a region initially scored as moderate potential and, through pattern matching against stress-response pathways, elevate its significance. The results of this transformation are detailed in Table 4.
The combined outputs from both the PSO and GA stages produce a comprehensive and actionable view of the processed data. As evidenced by the transition from Table 3 to Table 4, the method significantly improves the clarity of the genomic signal. Sample S004, for example, evolved from a raw score of 0.71 to an optimized score of 0.88, allowing for a high-confidence classification (95%) as a “Cognitive” performance marker. Similarly, the system successfully extracted a “Stress-Regulator” signal from the structurally complex Centromere region (S001), a task often prone to failure in standard linear analysis. This integrated model is therefore well-suited for large-scale deployment in military genomics, where the accuracy of individualized risk profiles and the reliability of detailed physiological measurements are critical for operational planning and soldier welfare.
Figure 1 illustrates the user interface for a hybrid system integrating particle swarm optimization and a genetic algorithm for genomic analysis. The upper panel displays the PSO settings: 30 particles, an inertia weight of 0.7, and a cognitive and social coeficient of 1.5. The lower panel presents GA parameters: crossover rate of 0.7, mutation rate of 0.15, elitist selection retaining the top 50% of candidates, and a noise interval of ±0.05 . In the “Gene Classification Mapping” section, optimized scores are linked to gene categories: scores above 0.7 indicate “Stress-Regulator,” 0.6–0.7 for “Immune-Response,” 0.5–0.6 for “Metabolic-Controller,” and below 0.4 for “Neural-Modulator.” Buttons labeled “Run Algorithm Again” and “Download Chart” enable re-executing the algorithm or exporting the results chart.
These settings serve as the default configuration, which can be adjusted for specific test datasets, for example, S001–S004. The PSO and GA parameters were selected to optimize the trade-of between computational eficiency and accuracy, a critical consideration in military genomics applications. The gene classification mapping translates optimized scores into predicted functional roles, facilitating the identification of genes linked to stress tolerance and physical performance. The subsequent step focuses on visualizing the outcomes to evaluate the algorithm’s overall efectiveness.
Figure 2 displays a comparative visualization of initial versus optimized potential scores for samples S001–S004. The bar chart uses light blue bars to represent the starting scores and green bars for the optimized results. Specifically, S001 increases from 0.45 to 0.63, S002 from 0.62 to 0.91, S003 from 0.38 to 0.46, and S004 from 0.71 to 0.91. These improvements illustrate that the hybrid algorithm efectively enhances gene detection across the tested samples.
The visualization demonstrates a noticeable enhancement in potential scores, especially for samples S002 and S004, which both reach 0.91 after optimization. This suggests that the algorithm successfully identifies previously hidden genes with significant functional relevance, including those involved in stress regulation or cognitive performance. The contrast between the original and optimized scores underscores the advantage of integrating PSO with GA, highlighting the approach’s promise for advanced genomic analyses in military applications.
The design and evaluation of a hybrid system integrating particle swarm optimization with a genetic algorithm has proven highly efective for uncovering concealed genes in military genomic maps. Testing on datasets S001–S004 revealed that the approach substantially enhances gene potential scores, particularly in structurally complex regions such as centromeres and segmental duplications. By combining PSO-driven broad scanning with GA-based refinement, the framework reliably detects genes associated with stress tolerance and key factors of physical performance for applications in military genomics.
The findings show that optimized scores of 0.91 for samples S002 and S004 surpass their initial values of 0.62 and 0.71, indicating efective identification of genes with high functional relevance. Comparative visualizations confirmed a consistent increase in scores, while gene classification allowed assignment to categories such as “Stress-Regulator” and “Cognitive.” The approach also demonstrated scalability, making it well-suited for analyzing extensive datasets that reflect the genetic variability among military personnel.
These results suggest that the hybrid PSO-GA algorithm has significant potential for personalized
medicine applications in the military [
The potential applications of this hybrid algorithm extend far beyond military genomics,
encompassing fields such as healthcare [
These estimates, derived from analogous technologies, illustrate the wide-ranging potential of the proposed method. In the medical field, applying the algorithm to 10,000 patient cases, resulting in a 35% improvement in accuracy, could prevent approximately 3,500 diagnostic errors annually, saving nearly 2,000 hours of analysis time. In agriculture, a 17.5% increase in crop yields across 50,000 hectares could produce an additional 8,750 tons of corn, translating into a 10–15% rise in farmer revenue. These figures underscore the value of tailoring the algorithm to diverse applications, ofering tangible, field-specific benefits.
Thanks to our colleagues for their valuable advice and constructive feedback, which helped refine the methodology and interpret the results.
The authors have not employed any Generative AI tools.