Unveiling facial kinship: The BioKinVis dataset for facial kinship verification and genetic association studies.

Facial image-based kinship verification represents a burgeoning frontier within the realms of computer vision and biomedicine. Recent genome-wide association studies have underscored the heritability of human facial morphology, revealing its predictability based on genetic information. These revelations form a robust foundation for advancing facial image-based kinship verification. Despite strides in computer vision, there remains a discernible gap between the biomedical and computer vision domains. Notably, the absence of family photo datasets established through biological paternity testing methods poses a significant challenge. This study addresses this gap by introducing the biological kinship visualization dataset, encompassing 5773 individuals from 2412 families with biologically confirmed kinship. Our analysis delves into the distribution and influencing factors of facial similarity among parent-child pairs, probing the potential association between forensic short tandem repeat polymorphisms and facial similarity. Additionally, we have developed a machine learning model for facial image-based kinship verification, achieving an accuracy of 0.80 in the dataset. To facilitate further exploration, we have established an online tool and database, accessible at http://120.55.161.230:88/.

Genome-wide association studies combined with k-fold cross-validation identify rs17822931 as an ancestry-informative marker in Han Chinese population.

DNA-based ancestry inference has long been a research hot spot in forensic science. The differentiation of Han Chinese population, such as the northern-to-southern substructure, would benefit forensic practice. In the present study, we enrolled participants from northern and southern China, each participant was genotyped at ∼400 K single-nucleotide polymorphisms (SNPs) and data of CHB and CHS from 1000 Genomes Project were used to perform genome-wide association analyses. Meanwhile, a new method combining genome-wide association study (GWAS) analyses with k-fold cross-validation in a small sample size was introduced. As a result, one SNP rs17822931 emerged with a p-value of 7.51E - 6. We also simulated a huge dataset to verify whether k-fold cross-validation could reduce the false-negative rate of GWAS. The identified ABCC11 rs17822931 has been reported to have allele frequencies varied with the geographical gradient distribution in humans. We also found a great difference in the allele frequency distributions of rs17822931 among five different cohorts of the Chinese population. In conclusion, our study demonstrated that even small-scale GWAS can also have potential to identify effective loci with implemented k-fold cross-validation method and shed light on the potential maker of rs17822931 in differentiating the north-to-south substructure of the Han Chinese population.

Evaluation of a SNP-STR haplotype panel for forensic genotype imputation.

Short tandem repeat polymorphism (STR)-based individual identification is a popular and reliable method in many forensic applications. However, STRs still frequently fail to find any matched records. In such cases, if known STRs could provide more information, it would be very helpful to solve specific problems. Genotype imputation has long been used in the study of single nucleotide polymorphisms (SNPs) and has recently been introduced into forensic fields. The idea is that, through a reference haplotype panel containing SNPs and STRs, we can obtain unknown genetic information through genotype imputation based on known STR or SNP genotypes. Several recent studies have already demonstrated this exciting idea, and a 1000 Genomes SNP-STR haplotype panel has also been released. To further study the performance of genotype imputation in forensic fields, we collected STR, microhaplotype (MH) and SNP array genotypes from Chinese Han population individuals and then performed genotype imputation analysis based on the released reference panel. As a result, the average locus imputation accuracy was ∼83 % (or ∼70 %) when SNPs in the SNP array (or MH SNPs) were imputed from STRs, and was ∼30 % when highly polymorphic markers (STRs and MHs) were imputed from each other. When STRs were imputed from SNP array, the average locus imputation accuracy increased to ∼48 %. After analyzing the match scores between real STRs and the STRs imputed from SNPs, ∼80 % of studied STR records can be connected to corresponding SNP records, which may help for individual identification. Our results indicate that genotype imputation has great potential for forensic applications.

Could routine forensic STR genotyping data leak personal phenotypic information?

The application of forensic genetic markers must comply with privacy rights and legal policies on a premise that the markers do not expose phenotypic information. The most widely-used short tandem repeats (STRs) are generally viewed as 'junk' DNA because most STRs are located in non-coding regions and therefore refrain from leaking phenotypic traits. But with a deepening understanding of phenotypes and underlying genetic structure, whether STRs could potentially reflect any phenotypic information may need re-examining. Therefore, we performed the following analyses. First, we analyzed the association between 15 STRs and three facial characteristics (single or double eyelid, with or without epicanthus, unattached or attached earlobe) on 721 unrelated Han Chinese individuals. Then, we collected 27199 individuals' STRs and geographic data from the literature to investigate the association between STRs and bio-geographic information, and predict geographic information by STRs on additional 1993 unrelated individuals. We found that there was scarcely any association between STRs with studied facial characteristics. Although allele19 in D2S1338 and allele 18 in FGA (P = 0.0032, P = 0.0030, respectively after Bonferroni correction) showed statistical significance, the prediction effectiveness was very low. For the STRs and bio-geographic information, the principal component analysis showed the first three components could explain 87.7% of the variance, but the prediction accuracy only reached 25.2%. We demonstrated that the forensic phenotypes are usually complex traits, it is hardly possible to uncover phenotypic information by testing only dozens of STR loci.

A highly polymorphic panel of 40-plex microhaplotypes for the Chinese Han population and its application in estimating the number of contributors in DNA mixtures.

Microhaplotypes (MHs) have great potential in multiple forensic applications and have proven to be promising markers in complex DNA mixture analysis. In this study, we developed a multiplex panel of 40 highly polymorphic MHs for the Chinese Han population, evaluated its forensic values, and explored its application in predicting the number of contributors (NOCs) in DNA mixtures. The panel consisted of 20 newly proposed loci and 20 previously reported loci with lengths spanning less than 120 bp. The average effective number of alleles (Ae) was 3.77, and the cumulative matching probability (CMP) and the cumulative power of exclusion (CPE) reached 1.2E-37 and 1-2.1E-12, respectively, in the Chinese Han population from the 1000 Genomes Project. Further validation on 150 Chinese Han individuals showed that Ae ranged from 2.62 to 4.41 with a mean value of 3.61, and CMP and CPE were 3.61E-36 and 1-1.84E-12, respectively, indicating that this panel was informative for personal identification and paternity testing in the studied population. To estimate NOC in DNA mixtures, we developed a machine learning model based on this panel. As a result, the accuracies in artificial DNA mixtures reached 95.24% for 2- to 4-person mixtures and 83.33% for 2- to 6-person mixtures. Furthermore, the NOC estimation on simulated profiles with allele dropout showed that this panel was still robust under slight dropout. In conclusion, this panel has value for forensic identification and NOC estimation of DNA mixtures.

Association between a Single Nucleotide Polymorphism in the 3'-UTR of ARHGEF18 and the Risk of Nonidiopathic Pulmonary Arterial Hypertension in Chinese Population.

ARHGEF18 has been identified as upregulated in the lung tissues of rat models of pulmonary artery hypertension introduced by hypoxia or monocrotaline (MCT). We used online SNP function prediction tools to screen the candidate SNPs that might be associated with the regulation of the ARHGEF18 expression. The result suggested that rs3745357 located in the 3'-untranslated region of ARHGEF18 is probably a genetic modifier in the process. In the present study, we aimed to investigate the association between ARHGEF18 rs3745357 polymorphism and nonidiopathic pulmonary arterial hypertension susceptibility (niPAH). A total of 293 participants were included in the case-control study (117 patients and 176 healthy controls). The rs3745357 variant was discriminated by using cleaved amplification polymorphism (CAP) sequence-tagged site technology. Although the overall allele and genotype frequencies of rs3745357 in niPAH patients were close to those of the control group, significant differences have been identified when we further divided the niPAH patients into subgroups with or without coronary heart disease (CHD). Rs3745357 C allele frequency was significantly higher in niPAH patients without CHD history (p = 0.001), while the frequency was significantly lower in niPAH patients with CHD history (p = 0.017) when compared to control subjects. The distribution of genotype frequencies was also quite different. After adjustment by gender and age, significant differences were found between patients with CHD history and controls. The results suggest that the ARHGEF18 rs3745357 variant may be used as a marker for the genetic susceptibility to niPAH.

Genetic polymorphism and evolutionary differentiation of Eastern Chinese Han: a comprehensive and comparative analysis on KIRs.

Killer cell immunoglobulin-like receptor genes, namely KIRs, cluster together within the 160 kb genomic DNA region. In this study, we used PCR-SSP approach and successfully identified the genotype of 17 KIR genes in 123 independent healthy donors residing in the Jiangsu province, China. All individuals were positive at the 7 genes. The observed carrier gene frequencies (OFs) of remaining 10 KIRs ranged from 14.63% (KIR2DS3) to 95.93% (KIR3DL1). We found 27 distinct genotypes excluding KIR1D. The most frequent occurred in 63 individuals (51.22%). The linkage disequilibrium analysis signified 29 positive and 6 negative relations in 45 pairwise comparisons. To study population differentiation, we drew a Heatmap based on the data of KIRs from 59 populations and conducted Hierarchical Clustering by Euclidean distances. We next validated our results by estimating pairwise DA distances and illustrating a Neighbor-Joining tree, as well as a MDS plot covering 3 additional Chinese Han groups. The phylogenetic reconstruction and cluster analysis strongly indicated a genetically close relationship between Eastern and Jilin Hans. In conclusion, the present study provided a meritorious resource of KIR genotyping for population genetics, and could be helpful to uncover the genetic mechanism of KIRs in immune disease in the future.