Application of BT and PC-BT in Homo sapiens gene prediction

Autor: Soma Barman, Antara Ghosh
Rok vydání: 2015
Předmět:
Zdroj: Microsystem Technologies. 22:2691-2705
ISSN: 1432-1858
0946-7076
DOI: 10.1007/s00542-015-2573-9
Popis: Digital signal processing is gaining popularity in gene classification and prediction. Spectral content of gene is estimated using Blackman---Tukey (BT) method for identification of Homo sapiens cancer genes. A data reduction method; principal component analysis (PCA) is used prior to Blackman---Tukey method (PC-BT) for better identification of genes. Selection of proper model order is prime important in PCA method for considering the number of principal components (PCs) and the performance of the method depends on optimal selection of PCs. Cumulative percent variance and scree test residual percent variance are used as measurement metric for selection of PCs. Quality factor is used to judge the performance of the estimators. Simulation results show the clarity of spectrum plot in PC-BT method is 200 % higher than BT method. The spectral peaks observed in cancer genes are used to screen out cancer genes from healthy genes. The methods are successfully tested on breast, prostate and colon Homo sapiens. Healthy and cancer Homo sapiens genes are downloaded from Cancer Genome Anatomy Project (CGAP) site and National Centre for Biotechnology Information (NCBI) site.
Databáze: OpenAIRE