Development of Accurate Relatedness Estimation for Malaria Vectors

Identification of close kin is a critical quality control step for genomic analysis, as well as an analytical goal in itself. Demographic processes such as inbreeding, and variation in recombination rate over small genomes, may make accurate relatedness inference challenging for Anopheles taxa. Kinship statistics derived from the condensed Jacquard coefficients offer the possibility of more accurate relatedness estimation, especially where inbreeding is an issue. However, standard tools for these calculations such as NGSRelate are difficult to integrate into Python and use a computationally expensive expectation-maximization algorithm. The recent development of Kindred suggests a constrained least-squares formulation that may be fast enough to be used in an interactive environment. This project will develop a Python library that combines both approaches in a hybrid algorithm that should make estimating the Jacquard coefficients both fast and accurate. This library will be integrated into the MalariaGEN API and make relatedness estimation of Anopheles taxa accessible to a broad audience on Google Colab.