Search bioRxiv⌕ Search

Biology subjects

Choi, K. Y.

Publications and source records attributed to Choi, K. Y..

2 recordsLinked to original sources

A Mathematical Methodology to Predict Phenotype from Genotype

From a mathematical perspective, a genome can be viewed as categorical data whose elements include genetic variants in SNPs. Then, the problem of predicting phenotype from genotype can be reduced to the problem of classifying categorical data. We will define a metric function at each SNPs position of an individual such that the arithmetic mean of a specific phenotype group is always less than or equal to that of the comparison group. Then, in light of a mathematical principle, calculating the sum of the distances(metrics) of some SNPs positions could clearly classify the two groups, when difficult to classify them with one specific position. Using this methodology, we found that even small differences in each SNPs combined could classify phenotypes ; combining 51 SNPs positions in autosomes made possible to distinguish males from females with great accuracy. We also found that East Asian people could be distinguished from other people with 100% accuracy, so we obtained a mathematical definition of East Asian people. The mathematical methodology classifies phenotypes with collecting small differences of genotypes throughout the entire genome, which could be useful for predicting diseases affected by the genome and be useful in pharmacogenomics.

bioinformatics↗

DeepParcellation: a novel deep learning method for robust brain magnetic resonance imaging parcellation in older East Asians

Accurate parcellation of cortical regions is crucial for distinguishing morphometric changes in aged brains, particularly in degenerative brain diseases. Normal aging and neurodegeneration precipitate brain structural changes, leading to distinct tissue contrast and shape in people aged > 60 years. Manual parcellation by trained radiologists can yield a highly accurate outline of the brain; however, analyzing large datasets is laborious and expensive. Alternatively, newly-developed computational models can quickly and accurately conduct brain parcellation, although thus far only for the brains of Caucasian individuals. DeepParcellation, our novel deep learning model for 3D magnetic resonance imaging (MRI) parcellation, was trained on 5,035 brains of older East Asians (Gwangju Alzheimers & Related Dementia) and 2,535 brains of Caucasians. We trained full 3D models for N-way individual regions of interest using memory reduction techniques. Our method showed the highest similarity and robust reliability among age-ethnicity groups, especially when parcellating the brains of older East Asians.

neuroscience↗