@2024 Afarand., IRAN
ISSN: 1735-7675 Kowsar Medical Journal 2010;15(1):23-29
ISSN: 1735-7675 Kowsar Medical Journal 2010;15(1):23-29
Application of near-infrared spectroscopy and support vector machine in detection of HIV-1 infection
ARTICLE INFO
Article Type
Original ResearchAuthors
Rahimi A. A. (1 )Bahmani M. Kh. (* )
Masoudnejad A. (2 )
Sa’adati M. (1 )
Khosravi A. (1 )
Shafi’khani A. (3 )
Mohammadzade J. (4 )
(* ) Applied Biotechnology Research Center, Baqiyatallah University of Medical Sciences, Tehran, Iran
(1 ) Biology Research Center, Faculty of Basic Sciences, Imam Hossein University, Tehran, Iran
(2 ) Biochemistry & Biophysics Research Center, Tehran University, Tehran, Iran
(3 ) Physics Research Center, Institute of Fundamental Knowledge, Tehran, Iran
(4 ) Faculty of Sciences, Branch of Karaj, Islamic Azad University , Karaj, Iran
Correspondence
Address:Phone:
Fax:
mkbahmani@sums.ac.ir
Article History
Received: February 13, 2008Accepted: September 10, 2008
ePublished:
ABSTRACT
Aims
This study was performed to investigate the possibility of diagnosis of HIV-1 human serum infection and to differentiate the infected serum from healthy ones using near infrared (NIR) spectroscopy along with Support Vector Machines (SVM) analysis as a rapid, cost-effective and non-destructive method.
Materials & Methods 35 HIV-1 infected serums and 15 healthy ones were selected and scanned within the range of 600-1100 nm using spectrophotometer device and their absorbance was obtained. The given results were then subjected to SVM for conducting the training phase and calibration process. After determination of training structure of SVM, absorbance data obtained from 21 HIV-1 infected and 20 uninfected serums at the same wavelength range were subjected to SVM as the unknown samples for test process and diagnosis of infection or the lack of it.
Results Among 21 infected unknown samples, 2 cases were diagnosed as uninfected and the rest were diagnosed as infected and from 20 uninfected unknown samples, only 1 case was diagnosed by SVM as infected. The sensitivity of the method was estimated 91% and its specificity was 95%.
Conclusion NIR-spectroscopy along with SVM analysis can be used as a pre-diagnostic method for detection of human serum infection with HIV-1.
Materials & Methods 35 HIV-1 infected serums and 15 healthy ones were selected and scanned within the range of 600-1100 nm using spectrophotometer device and their absorbance was obtained. The given results were then subjected to SVM for conducting the training phase and calibration process. After determination of training structure of SVM, absorbance data obtained from 21 HIV-1 infected and 20 uninfected serums at the same wavelength range were subjected to SVM as the unknown samples for test process and diagnosis of infection or the lack of it.
Results Among 21 infected unknown samples, 2 cases were diagnosed as uninfected and the rest were diagnosed as infected and from 20 uninfected unknown samples, only 1 case was diagnosed by SVM as infected. The sensitivity of the method was estimated 91% and its specificity was 95%.
Conclusion NIR-spectroscopy along with SVM analysis can be used as a pre-diagnostic method for detection of human serum infection with HIV-1.
CITATION LINKS