The Research Of Gene Identification Questions Based On Mathematical Statistics

The principal means of gene identification is experiments which based on the living cells or organism. We can obtain the gene sequence in chromosome by a number of different genes of homologous recombination rate of statistical analysis. If a large number of similar analyses can be made, we can determine the approximate location of every gene. However, a huge quantity of genomic information has been acquired by human; the method which relayed on the slower experimental analysis cannot meet the needs of the gene identification. With the development of computer technology, adopt the computer algorithms to identify genes has gradually become the primary means.When facing the large number of complex gene sequence data; how to better and more quickly get the accurate genetic information, how to calculate the power spectrum and the signal to noise ratio successfully in a lot of gene sequence data; how to quickly achieve the gene identification, all above are important subject before us which have a research significance. The paper makes the extraction for gene feature and identification, builds the model of gene feature extraction and identification, finally, implements by means of computer.Firstly, the paper adopt the Voss mapping and Z-curve mapping to make the abstract symbol sequence map into a numerical sequence through the method of signal processing, and drawn the spectrogram, constructed the fast calculation method of power spectrum and signal-to-noise ratio; after deducing, got the conclusion that " the power spectrum and the signal to noise ratio obtained by Z-curve mapping and Voss genetic data has multiple relationship", and its power spectrum coefficient P is4, the proportion of the signal-to-noise ratio R is4/3; This paper has also established the gene identification model which based on spectrum and SNR, and applied to the research of gene mutation data, and has been obtained the corresponding conclusions.