2024 : 7 : 16
omid Jazayeri

omid Jazayeri

Academic rank: Assistant Professor
ORCID:
Education: PhD.
ScopusId:
Faculty: Science
Address:
Phone: 011-35302450

Research

Title
AICRF: ancestry inference of admixed population with deep conditional random field
Type
JournalPaper
Keywords
admixed haplotype; admixed population; ancestry inference; proper window length; local ancestry.
Year
2023
Journal Journal of Genetics
DOI
Researchers FARHAD ALIZADEH ، Hamid Jazayeriy ، omid Jazayeri ، FATEMEH VAFAEE

Abstract

Ancestry inference of admixed populations is an important issue in anthropology and studies of gene discovery, and characterization. Usually, local ancestor inference (LAI) methods use fixed-length windows to divide chromosomes into smaller blocks. The accuracy of LAI algorithms will decrease if a window with an inappropriate length is used to infer the ancestry of admixed individuals. In this study, we first present a heuristic function to determine a proper window length for LAI methods. This heuristic is based on the distance between the ancestral populations of admixed individuals. Then we introduce a method for ancestry inference of admixed population with deep conditional random field (AICRF). AICRF uses a conditional random field (CRF) parameterized by probable extreme learning machines (PELMs) trained on reference panels where PELM is a novel probabilistic ELM classifier. This method does not require many statistical or biological parameters. We evaluate the performance of AICRF in comparison with RFMix. Experimental results show that AICRF is more accurate than RFMix with increasing admixture times.