Ryusei Shingaki

A proposal for improvement of genotyping performance for ethnically homogeneous population using DNA microarray

DNA microarray is used to determine the genotypes of several hundred thousand to several million SNPs (Single Nucleotide Polymorphisms) on multiple samples at a time. In the conventional method of genotyping using DNA microarray, it is assumed that each SNP has three types of genotypes: two homozygous and one heterozygous genotypes. However, in an ethnically homogeneous population, there are cases when all the samples of a SNP belong to one homozygous genotype, and there are some other cases, especially in the SNPs of low MAFs (Minor Allele Frequencies), each sample belongs to either of the two genotypes: one homozygous and one heterozygous genotypes. In those cases, the conventional method of genotyping may fail to properly determine the genotypes of the samples. In this paper, we propose a new genotyping method, which can be used as a post-processing technique of the conventional genotyping method, for re-judgment of the SNPs having one or two types of genotypes. The proposed method takes fluctuations of the fluorescence intensities of the signals of DNA microarray into account, assigns genotypes to samples from those genotype patterns that may occur under natural mating conditions and applies different genotype judgment methods depending on the number of genotype clusters of a SNP. We evaluate our proposed method using the data of 1000 Genome Project and have found that our proposed method is able to improve the genotyping performance of the conventional method.

Resident-presence/absence estimation by unsupervised threshold learning for home energy consumption and its application to resident profiling

We present a new method to estimate resident presence/absence status of a home using only home total power consumption data. The method is unsupervised in that the threshold computation for estimation is conducted without training data of presence/absence. From the estimation records collected for a certain period, we also automatically analyze a lifestyle profile of the residents living in the home.