|Institution:||University of Oslo|
|Full text PDF:||https://www.duo.uio.no/handle/10852/9568|
Microsatellites are widely used genetic markers. Although there will often be missing values in a set of microsatellite data, little work has been done to find a good method to estimate these missing values. K-Nearest Neighbours based methods have been shown to do well with other types of data, and this method was tried out for microsatellite data as well. The results show that the success of this method in estimating missing values in microsatellite data is dependent on the data set. Although well suited for some data, care should be taken when applying the method to very small data sets or data sets where a lot of the individuals are closely related.