Heuristic Analysis of Genomic Sequence Processing Models for High Efficiency Prediction: A Statistical Perspective.

Area of Science:

• Genomics and Bioinformatics
• Computational Biology
• Machine Learning in Biology

Background:

• Genome sequences contain diverse characteristics like species type, genotype, and disease markers.
• Various deep learning models (CNNs, DBNs, MLPs) are used for genome analysis but differ in performance and application.
• Selecting the optimal genome processing model is challenging due to algorithmic variations.

Purpose of the Study:

• To review and compare various deep learning and machine learning models for genome sequence analysis.
• To facilitate model selection for researchers by evaluating performance metrics.
• To provide insights into the efficient application of different genomic processing models.

Main Methods:

• Comparative analysis of deep learning and machine learning models for genomic data.
• Evaluation of models based on accuracy, precision, recall, computational complexity, and processing delay.
• Review of specific models including Repeated Incremental Pruning to Produce Error Reduction with Support Vector Machine (Ripper SVM), CNN Bayesian method, and Bidirectional Long Short-Term Memory with CNN (BiLSTM CNN).

Main Results:

• Ripper SVM achieved 99.7% accuracy for multiple genomic data.
• CNN Bayesian method showed 99.27% accuracy for cancer genomic data.
• BiLSTM CNN demonstrated the highest accuracy of 99.95% for Covid genome analysis.

Conclusions:

• Different genomic processing models exhibit varying accuracies for specific applications.
• The review provides a comparative analysis of model performance, including precision and recall.
• Recommendations are made for the efficient use of these models in genomic research.