Development of a forensic skin colour predictive test

作者：Maronas Olalla^*; Phillips Chris; Soechtig Jens; Gomez Tato Antonio; Cruz Raquel; Alvarez Dios Jose; Casares de Cal Maria; Ruiz Yarimar; Fondevila Manuel; Carracedo Angel; Lareu Maria V

来源：Forensic Science International-Genetics, 2014, 13: 34-44.

DOI：10.1016/j.fsigen.2014.06.017

• Skin colour prediction
• Pigmentation
• Forensic phenotyping
• Externally visible traits
• Naive Bayes classification
• Linear regression analysis

There is growing interest in skin colour prediction in the forensic field. However, a lack of consensus approaches for recording skin colour phenotype plus the complicating factors of epistatic effects, environmental influences such as exposure to the sun and unidentified genetic variants, present difficulties for the development of a forensic skin colour predictive test centred on the most strongly associated SNPs. Previous studies have analysed skin colour variation in single unadmixed population groups, including South Asians (Stokowski et al., 2007, Am. J. Hum. Genet, 81: 1119-32) and Europeans (Jacobs et al., 2013, Hum Genet. 132: 147-58). Nevertheless, a major challenge lies in the analysis of skin colour in admixed individuals, where co-ancestry proportions do not necessarily dictate any one person%26apos;s skin colour. Our study sought to analyse genetic differences between African, European and admixed African-European subjects where direct spectrometric measurements and photographs of skin colour were made in parallel. We identified strong associations to skin colour variation in the subjects studied from a pigmentation SNP discovery panel of 59 markers and developed a forensic online classifier based on naive Bayes analysis of the SNP profiles made. A skin colour predictive test is described using the ten most strongly associated SNPs in 8 genes linked to skin pigmentation variation.

• 出版日期2014-11
• 单位司法部司法鉴定科学技术研究所