Computational Intelligence Strategies Throughout Bioengineering

Aminata Garba^*
*Correspondence: Aminata Garba, Department of Computer Engineering, Ecole Polytechnique Federale de Lausanne, Rte Cantonale, Lausanne, Switzerland, Email:

Introduction

Applications in systems biology, pharmacogenomics, epigenetics, genomics, and proteomics have risen significantly over the past several years, which has led to a massive increase in the production of complex, highly dimensional data. Bioinformatics areas constantly use high-dimension, small sample data. Genome-wide studies produce a significant amount of data, necessitating the use of soft computing methods (SCMs) that can handle this volume of data, such as artificial neural networks, fuzzy systems, evolutionary algorithms, metaheuristic and swarm intelligence algorithms, statistical model algorithms, etc. Many more bioinformatics applications now make use of soft computing techniques. It is a key technique for examining the biological molecules' interactions and underlying processes in many disorders. Understanding the fundamental causes of human sickness problems requires the use of genetic evidence processing methods. A significant expansion in the use of genetic capabilities has led to the creation of a vast amount of complex, multidimensional data. This is because they can handle multi-dimensional, complex datasets like those produced by DNA microarray and protein mass spectrometry investigations. Neural networks have however been enough to authenticate biomarkers and diagnose medical issues. In order to avoid overfitting, develop more effective classifiers, and provide deeper understanding of the underlying causal links, feature selection is employed in conjunction with classifier design [1,2].

Description

SCMs may also be used in the evolution domain, particularly in the reconstruction of phylogenetic trees. A phylogenetic tree is a graphical illustration of an organism's evolutionary connections, which are deduced from similarities and differences in the physical and genetic characteristics of various biological species. The alignment of biological sequences is caused by a variety of distinct factors. Discovering the structural and functional similarities of sequences in biology is facilitated by sequence alignment. These aligned sequences are used by scientists to create phylogenetic trees, describe protein families, and predict protein structure. Generally, they were made up of criteria like physical characteristics, metabolic characteristics, etc., but now that there are so many genome sequences accessible, phylogenetic tree construction techniques are based on comparing various genomes. Gene selection, clustering and classification, signal processing, and image analysis are just a few of the bioinformatics issues that SCMs are ideally suited for. In bioinformatics tasks, supervised or unsupervised classification with multi-dimensional input variables is regularly encountered. SCMs can handle complex multidimensional datasets as a result. SCMs are employed to address several bioinformatics issues. SCMs fall into two categories: supervised learning rules and unsupervised learning rules [3,4].

In order to clarify links among sample groupings, unsupervised or clustering approaches are employed to group comparable genomic or proteomic profiles. Based on the expression characteristics of the biomarkers across patient samples, these approaches also allocated biomarkers to subgroups. Clustering is delimited since it is unable to take expert knowledge into account, although being effective for exploratory investigation. Soft computing techniques (SCMs) entail programming computers to enhance a performance standard using test data or prior knowledge. In essence, soft computing is a heterogeneous set of ideas and techniques. Instead, it is a collaboration of several approaches that in some way or another adhere to its guiding idea. Currently, the goal of soft computing is to achieve tractability, resilience, and inexpensive solution costs by making use of the tolerance for uncertainty and imprecision. In order to attain practicability, resilience, and cheap solution costs, SCMs deal with uncertainty, approximation, partial truth, and imprecision. It should be noted that the effectiveness of learning and inference algorithms might be just as crucial as their space and temporal complexity, transparency, and interpretability. Artificial Neural Networks (ANN) are information processing models that are implemented in hardware or software and are based on scientifically researched biological brain processes. Artificial neural networks have the capacity to uncover patterns and discern trends from ambiguous or complex data that are difficult for humans or other computer approaches to easily understand [5].

In various scientific and technical fields, including biomedicine and bioinformatics, artificial neural networks (ANN) have mostly been utilised to study the complex interactions between input and output variables. The majority of ANN algorithms used to handle bioinformatics issues include Back- Propagation (BP), Radial Based Function (RBF), and Support Vector Machines (SVM). In order to solve a binary classification task under supervision, Support Vector Machines (SVM) are utilised. The learning problem is framed as a quadratic optimization problem, and the error surface is free of any local minima and has a global optimum. The goal of SVM is to construct an ideal separating hyper plane that maximises the distance between two classes. SVM implement this desired quality on the basis of the structural risk reduction concept. The knowledge-based model with linguistic rules is called the fuzzy system model. For each input, output, and set of rules, fuzzy sets are specified. The ability to analyse this information and determine output values for given input data is ensured by fuzzy logic [2,4].

Conclusion

For the purpose of obtaining genotypes and collecting samples of patients and controls, bioinformatics data must be used. It makes sense to say that significant effort should be put into making sure that the analytic methods used on the data acquired are as successful as feasible. Because a biologist might not be able to comprehend the results of soft computing approaches. Someone who is knowledgeable on parallel computing, for instance, would be needed to help with an issue that has to be answered by several processors. The programme must be simple to use and produce output that is both visible and navigable for it to be intuitive to a biologist. Therefore, we may conclude that biologists must work with soft computing techniques in order to more effectively handle bioinformatics difficulties.

Acknowledgement

None.

Conflict of Interest

The authors reported no potential conflict of interest.

References

1. Guerra-Montenegro, Juan, Javier Sanchez-Medina, Ibai Lana and Javier Del Ser, et al. "Computational Intelligence in the hospitality industry: A systematic literature review and a prospect of challenges." App Soft Com 102 (2021): 107082.

Google Scholar, Crossref, Indexed at

2. Castillo, Oscar, Patricia Melin, Oscar Montiel Ross and Witold Pedrycz, et al. "Theoretical advances and applications of fuzzy logic and soft computing." (2007).

   Google Scholar, Indexed at

3. Zhang, Wengang, Runhong Zhang, Chongzhi Wu and Hanlong Liu. "State-of-the-art review of soft computing applications in underground excavations." Geosci Front 11 (2020): 1095-1106.

Google Scholar, Crossref, Indexed at

4. Xie, Hong-Bo, Tianruo Guo, Siwei Bai and Socrates Dokos. "Hybrid soft computing systems for electromyographic signals analysis: a review." Biomed Eng On 13 (2014): 1-19.

   Google Scholar, Crossref, Indexed at

5. Omolaye, Philip O., Joseph M. Mom and Gabriel A. Igwue. "A holistic review of soft computing techniques." App Comp Math 6 (2017): 93-110.

   Google Scholar, Crossref, Indexed at