BIOMEDICAL RESEARCH AND REVIEWS

ISSN 2631-3944

Deep Learning Augmented with Hybrid Non-linear Cellular Automata Platform for Addressing Open Problems in Bioinformatics

Pokkuluri Kiran Sree1 *, Usha Devi N2

1Professor, Department of CSE, Shri Vishnu Engineering College for Women, Bhimavaram, India
2Assistant Professor, Department of CSE, University College of Engineering, JNTU Kakinada, India

CitationCitation COPIED

Sree PK, Usha DN. Deep Learning Augmented with Hybrid Non-linear Cellular Automata Platform for Addressing Open Problems in Bioinformatics. Biomed Res Rev. 2020 Apr;3(3):126.

© 2020 Sree PK, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 international License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Abstract and Introduction

Most of the grand challenges in Bioinformatics can be addressed my Deep Learning and Cellular Automata [1]. We have developed various classifiers to address essential problems in Bioinformatics. After that, we have done a literature survey, which indicates that many computer algorithms are used to solve open issues like Protein Coding Region (PCR), Promoter prediction (PR), and Chronic Disease Prediction (CDP) individually [2,3]. After thorough research, we conclude that many of the chronic diseases, homology searches, gene finding, and genomic sequence finding could be addressed quickly with a combination of deep learning augmented with cellular automata [4-6]. 

We understood when a deep learning model is augmented with non-linear cellular automata, and we can have the following advantages.

  1. These give the best performance in several domains Bioinformatics in particular
  2. This combination reduces the purpose of feature engineering, as this is the most typical part of most of the classifiers.
  3. The architectures designed can be easily adapted to solve many problems also as we can make a single platform to address many issues.
  4. The augmentation of NLCA
  1. Gives more dynamism making the training less expensive.
  2. Learning quickly with fewer examples.
  3. Understanding of theoretical foundations.
  4. Processing with fewer hyper parameters
  5. Easy to understand the process as to what is happening and what is going to happen next

Keywords

Deep learning ; Bioinformatics ; Cellular automata ; Non-linear rules

Abbreviations

PCR: Protein Coding Regions

CDP: Chronic Disease Prediction

NLCA: None Linear Cellular Automata

CA: Cellular Automata

DNN: Deep Neural Network CNN:

Convolution Neural Network NL-MACA:

Non-Linear Multiple Attractor Cellular Automata

System Design

Based on the above intuition, the design applicable for various problems was given below in (Figure1). This system design is vital in addressing these problems as the input can be any sequence. Hybrid non-linear CA [7] rules are used to process the data either in the number of three or six or nine.

Figure 1: Design of a common platform to address various problems in Bioinformatics

PCR [8] prediction process the DNA sequence in terms of three as Codons are identified in three in number, and so on. A deep learning technique DNN (Deep Neural Network) [9] is used to classify the sequence into many classes from which once classes are predicted by NL-MACA (Non-Linear Multiple Attractor Cellular Automata). The classes or the prediction analysis can be the result of the entire design.

Conclusion

We have done a thorough literature survey to find a common platform to address various problems in Bioinformatics and are successful in this regard. The strength of the classifier lies in the adaptability, less overhead, more sensitivity, specificity, more accuracy, less computing time, and processing large datasets [10]. Complex problems like secondary, quaternary protein structure prediction can be carried within 0.7 Nanoseconds. 

References

  1. Bergeron BP. Bioinformatics computing. Prentice-Hall Professional. 2002 Nov. (Ref)
  2. Kanhere A, Bansal M. A novel method for prokaryotic promoter prediction based on DNA stability. BMC Bioinformatics. 2005 Jan;6(1).  (Ref)
  3. Keedwell E, Narayanan A. Intelligent bioinformatics: The application of artificial intelligence techniques to bioinformatics problems. John Wiley & Sons. 2005 May.   (Ref)
  4. Sree PK, Babu IR, Devi NU. Investigating an Artificial Immune System to strengthen protein structure prediction and protein coding region identification using the Cellular Automata classifier.   (Ref)International journal of bioinformatics research and applications. 2009 Oct;5(6):647-662.  (Ref)
  5. Sree PK, Babu IR, Devi NU. Identification of Promoter Region in Genomic DNA Using Cellular Automata Based Text Clustering. Int Arab J Inf Technol. 2010 Jan;7(1):75-78.  (Ref)
  6. Sree PK, Babu IR, Devi NU. PSMACA: An automated protein structure prediction using MACA (multiple attractor cellular automata). Journal of Bioinformatics and Intelligent Control. 2010;2(3):211-215.  (Ref)
  7. Baxevanis AD, Bader G, Wishart D. Bioinformatics. John Wiley & Sons. 2020.  (Ref)
  8. McGuffin LJ, Bryson K, Jones DT. The PSIPRED protein structure prediction server. Bioinformatics. 2000 Apr;16(4):404-405.  (Ref)
  9. Sree PK, Devi NU. Internet of things with Deep Learning for Societal Applications. 2018 Jan.  (Ref)
  10. Sree PK. A Novel Cellular Automata Classifier for COVID-19 Prediction. Journal of Health Sciences. 2020.  (Ref)