CORRECTION OF FLUORESCENT SIGNALS DISTORTED BY OVERLAPPING SPECTRAL BANDS IN THE "NANOFOR SPS" SEQUENCER
Автор: V. V. Manoilov, A. G. Borodinov, A. Ya. Loginov, A. I. Petrov, I. V. Zarutsky, V. E. Kurochkin
Журнал: Научное приборостроение @nauchnoe-priborostroenie
Рубрика: Приборостроение физико-химической биологии
Статья в выпуске: 2, 2025 года.
Бесплатный доступ
An important stage in the technology of massively parallel sequencing of nucleic acids is the process of measuring fluorescence signals. The spectral composition of the fluorescence signal carries information about the type of the fluorescence object (nucleotide) included in the DNA sample analyzed by the sequencer. The more nucleotides are correctly recognized during the reading process, the more accurate the result of reading the entire genome will be. The phenomenon of overlapping spectral components of the fluorescence signals of nucleotides A, C, G and T (cross-talk) is one of the negative factors that prevents reliable reading of the genome. To combat this phenomenon, a color correction algorithm is needed. This algorithm is decisive in determining the type of nucleotide located in these coordinates (cluster) in a given reading cycle. The paper considers a color correction algorithm that allows one to determine a crosstalk matrix using the linear regression method. Based on this matrix, an inverse matrix is determined, the use of which allows one to significantly reduce the influence of spectral band overlap on the accuracy of fluorescence signal measurement evaluation. The algorithm has been tested on experimental data and has shown high reliability of the results.
Nucleic acid sequencing, mathematical data processing, crosstalk phenomenon, crosstalk matrix, linear regression, least squares method nucleic acid sequencing, mathematical processing and classification of multivariate data, principal component analysis, machine learning
Короткий адрес: https://sciup.org/142244851
IDR: 142244851