Peter Kimathi

Base

Full name	Peter Kimathi
Summary	Improving genetic methods and algorithms to suit diverse population data
Type of researcher	Researcher (Professional)
Introduce yourself, your experience and your credentials	Hello, my name is Peter Kimathi and I’m from Kenya. I have a Bachelors Degree in Mathematics and in Computer Science from Maseno University. I also have a Masters Degree in Mathematical Sciences from the African Institute of Mathematical Sciences in Tanzania and also a Masters Degree in Human Genetics from the University of Cape Town in South Africa. Currently, I’m working as a bioinformatics analyst at the Centre for Proteomic and Genomic Research in South Africa.
Describe your research	Some of my roles at CPGR involve the statistical design and analysis of projects from different platforms, developing and implementing workflows or pipelines for the analysis of genomics data across platforms aligned with the ISO standards. For example, we implemented a computational microbiome analysis workflow in water to identify the microbiome communities present in drinking water, their abundance and classification. This type of analysis has a wide range of applications. For example, it can be used to detect the presence or absence of pathogenic bacteria in drinking water and it could also be applied in limiting any uncontrolled waste in the distribution of water. Several software packages exist for this kind of analysis. However, quantitative insight into microbial ecology, also known as CHIME is the most popular tool. CHIME implements a number of algorithm tools and our role at CPGR is to evaluate and implement in our pipeline the algorithm tools that are well suite for our analysis and make the workflow more reproducible and flexible.
Explain its significance	In terms of research interest, it’s a fact that the amount of variability explained by genome wide association studies, also known as the GWAS, remains very low for most of the complex traits. This implies that there is still much that is still not known as far as the genetics basis for most of the complex traits are concerned. On the other hand, genotype imputations, which estimate the missing genotypes from the genotypic data and polygenic risk analysis, also known as PRS, which quantifies the genetic risk of every individual in the study are among the popular techniques that have been proposed to help in uncovering the remaining portion of the missing heritability. However, there has been issues with this, especially when applied to diverse population data like Africans. For example, in one of our recent evaluation, we showed that even the recent and popular imputation tools still suffer when applied to African population data. Also, we assessed the performance of several polygenic risk methods in national population and we observed very low prediction values. As a mathematician and a software developer, my research interest is to explore ways in which these methods and algorithms can be improved or customized to suit the custom population data like of Africans.