Fraudulence Detection of Genetic Data

Abstract

We have seen a huge number of efforts to decipher human genome in past two decades such Human Genome Project. Such large projects push on the development of genetic testing services. Nowadays genetic testing can be used in various aspects such as diagnose disease, identify mutations, advise on treatment and detect carrier. Besides, a huge market for genetic testing has been formed. However, fraudulence of genetic data has been a problem and several samples of fraudulence in genetic data are noticed. Fraudulent data can lead very serious problems. For example, if fraudulent data are used in paternity testing and cancer testing, the result may be that a family is broke and one's life is destroyed. Therefore, the problem of genetic fraudulence should be prevented. The main cause of genetic fraudulence is simulator. Since most of DNA are shared by population, simulators can use reference and some profiles to simulate data which are very like real data from sequencing. Thus, one system, which can detect simulated data is required. However, there is no relevant product so far and even few related algorithms can be found. That is why this system is designed. The input for the system is original read file and output is a related detection report. Through finding features from a genetic data and analysis them, the system can give a grade to show whether the sample is simulated or not. Besides, a set of web pages are used to show related result for one detection. This report is used to give detail of this system. Firstly, an introduction is given. Some background is used to define the problem and show the value for the product. After that, some methodologies are introduced. Then, testing results will be showed to valid the efficiency of this system and discuss is given. In the end, a conclusion will introduce extra findings and future improvements.