FTG is a web server for analyzing nucleotide sequences to predict the genes
using Fourier transform techniques. This server implements the existing Fourier
transform algorithms for gene prediction and allows the rapid visualization
of analysis by output in GIF format. This server implements the following
algorithms to locate probable gene in sequence.
a) GENESCAN algorithm that allow detection of three-base periodicity and coding
region in genomic DNA.
b) LENGTHEN-SHUFFLE algorithm that allow improvement of the accuracy of gene
prediction in a short sequence or limited length of window by amplifying the
periodicity of three.
c) FTG algorithm, described in this report, that combines the above two algorithm.
A distinctive feature of the protein-coding genes is the existence of short-
range correlations in the nucleotide arrangement, the most prominent being the
1/3 periodicity. Fast Fourier Transformation (FFT) is a powerful technique that
is commonly used for detecting periodicity, patterns and tandem repeats in DNA
sequences (Pasquier et al., 1998). Fickett and Tung (1992) describe the signi-
-ficance of Fourier analysis for identification of protein-coding genes in DNA
sequence. FFT has the advantage over the other methods like Markov models,
neural networks and homology-based methods, of not requiring a learning set,
making it possible to identify a protein-coding gene with no known homologue.
Our aim is to provide FFT-based web server which implements existing Fourier
techniques for analyzing and detecting periodicity in DNA sequences in order
to assist the users in locating the protein-coding gene in their sequence.
If you use FTG please cite the following paper
Issac, B., Singh H., Kaur, H. and Raghava, G. P. S.(2002)
Locating probable genes using fourier transform approach.
Bioinformatics 18(1): 196-197.