A Novel Human STR Similarity Method using Cascade Statistical Fuzzy Rules with Tribal Information Inference

M. Rahmat Widyanto (University of Indonesia)
Reggio N. Hartono (University of Indonesia)
Nurtami Soedarsono (University of Indonesia)

International Journal of Electrical and Computer Engineering
Vol 6, No 6: December 2016

A novel human STR (Short Tandem Repeat) similarity method using cascade statistical fuzzy rules with tribal information inference is proposed. The proposed method consists of two cascade Fuzzy Inference Systems (FIS). The first FIS is to discriminate the tribal similarity, and the second FIS is to calculate the STR similarity. By using the allele marker’s statistical distribution probability density function as the membership function in the Fuzzy Rules of the first FIS, the new method makes it possible to tell the tribal similarity between two STR profiles. A 727 data acquired from tribal groups of Indonesia is used to examine the method produced promising result, being able to indicate higher tribal similarity score within a tribal group and lower similarity between tribal groups. In the light of Indonesia’s diverse tribal groups, these properties are able to be leveraged as a new way to improve the versatility of existing DNA matching algorithm.