DNA Compression Using Soft Computing

Abstract

The compression is a technique to represent data in another format which requires a smaller number of bits than original one Previously compression was done to reduce the size of data for less storage in the secondary storage but in the present scenario space is not a big issue But bandwidth is still limited to communicate data over the internet The size of genomic repositories become large in the present scenario Some efficient algorithms are required to compress large sequences for easy storage and communication A variety of standard text compression algorithms were applied to compress biological sequences but none found suitable In DNA there are four different kinds of nucleotides A Adenine G Guanine T Thymine and C Cytosine There are approximately three billion bases in a human DNA out of which 99% are equal in all humans GenBank which is a repository of biological sequences and a part of International Nucleotide Sequence Databases INSD observed that its size become approximately double in every period of 18 months In this research the author uses two soft computing techniques particle swarm optimization PSO and modified firefly MFF to compress DNA sequences In both the approaches DNA sequences are compressed using auto regressive modelling but in the first technique the model parameters are identified using PSO technique and in second technique these parameters are recognized using MFF algorithm During the literature survey it was found that the researchers had achieved the compression ratio approximately 1 point 69 bit per byte The average compression ratio achieved by the first proposed algorithm DCPSO is 1 point 42 bpb and by the second proposed algorithm HOARDNAComp is 1 point 39 bpb which is an important improvement over existing DNA sequence compression algorithm In the future the time and space complexities of proposed algorithms can be minimize using some efficient data structures and techniques newline