Forensic Application Research Of NGS-STR System In 2nd-degree Kinship Identification

Objective:Recently,the requests of kinship identification in addition to the conventional paternity test cases,such as grandparents&grandchildren,uncles&nephews,and half-sibs,are increasing with the widespread application of genetic markers,in the process of daily forensic practice.Such identification could have important application value for specific cases,such as unidentified remains of the victims or missing persons and the improvement of the search capabilities of criminal suspects.However,the identification of these complicated kinship relationships has not been effectively resolved,and increasing the number of genetic markers tested to obtain more information could be the first step to such solution.In the earlier stage of this research group,next generation sequencing（NGS）was used to construct a complex typing system containing amelogenin gene（Amel）and 42 autosomal short tandem repeats（STR）loci,which are still the mainstream genetic markers for human identification in forensic medicine.This study intends to evaluate the systematic application value of this typing system in the determination of between 2nd-degree kinship,in order to provide basic data for the studies of NGS-STR typing technology in the identification of complex kinship relationships,and a new technical solution in difficult cases such as the identification of the 2nd-degree kinship.Methods:1.Collection and verification of 2nd-degree kinship samplesSamples in 10 pedigrees came from the pedigree blood cards collected in the earlier stage of the research group were used in this study.The study was approved by the Medical Ethics Committee of Hebei Medical University and all subjects signed informed consent forms.DNA was extracted from blood cards using EZNA DNA Blood Midi kit,and quantified using Nano-Q^TMprotein nucleic acid quantifier.STR loci were amplified by PCR using Goldeneye ^TM20A,Goldeneye ^TM22NC and Microreader ^TM23sp kits.And then capillary electrophoresis was performed using ABI 3500 genetic analyzer to obtain the accurate genotyping at these loci.Paternity index or IBS score were calculated to verify the genetic relationships between the samples and to screening all the individual pairs with 2nd-degree relationship,i.e.,grandparent-grandchild,uncle-nephew or half-sibs.2.Detection of STR genotyping of family samples applying NGS-STR systemSamples within 2nd-degree relationship pairs were sequenced using the NGS-STR typing system constructed in the earlier stage of the research group,based on the Illumina MiSeq FGx ^TMplatform.And the sequenced data was analysed with parameters such as depth of coverage of sample（Do C of sample）,average depth of coverage of locus（Do C of locus）,locus sequence composition ratio and allele coverage ratio（ACR）.The results of NGS-STR were compared with those of CE-STR typing,and the differences between the two typing methods were analysed to evaluate the accuracy of the NGS-STR typing system and its consistency with the CE-STR technology.3.Threshold definition and efficacy evaluation of NGS-STR system in2nd-degree kinship identification.Cumulative likelihood ratio（CLR）and identity by state score（IBS）were calculated for the confirmed 2nd-degree relationship pairs and equal number of unrelated pairs randomly selected in the samples sequenced with the NGS-STR system.The distribution of such parameters in the two types of individual pairs were analysed.And then thresholds distinguishing 2nd-degree relationship with unrelated ones were set according to such analyses and evaluated with sensitivity and specificity.Results:1.Evaluation of sequencing data quality:（1）The fragment quality of constructed library was tested by Lab Chip^?GX Touch24.There were neither small fragment adapter peaks nor large fragment tailing peaks found in the library,which is consistent with the expected peak pattern of the library quality inspection.（2）The 7500 real-time quantitative results showed that the CT value of the blank control well was greater than 29,the slope of the standard curve was in the range of-3.1 to-3.5,the standard deviation between the replicate wells was less than 0.4,and the amplification efficiency was in the range of 90%to110%,indicating that the library quality inspection is qualified.（3）In this experiment,four times of pair-end（PE）PE300 sequencing with the MiSeq FGx ^TMMicro chip were performed on the Miseq FGx ^TMsequencing platform in Research Use Only Run（RUO）mode.The average values of main quality control indicators of offline data,i.e.,cluster density,clusters passing filter and base quality score Q30,were 1348.25 K/mm²,87.58%and 91.5%,respectively,which were all in line with Illumina’s official confirmation of the availability of sequencing data;（4）Do C analysis:The highest and lowest Do C of sample were 1065184×and 2665×,respectively,and the average Do C of all samples was147207±70720×（mean±SD）.The lowest average Do C of locus is D20S470,with a value of 1730±2030×（mean±SD）,and the highest average Do C of locus is TH01,with a value of 9866±10562×（mean±SD）.Locus with the largest dispersion of average Do C,i.e.,the worst stability,was D21S11,and the opposite locus was D2S441.（5）Analysis of the composition ratio of locus sequence:The analysis threshold was defined as 5 reads,10 reads,20 reads,30 reads,40 reads and5%of the total data at a single locus.Then the number of Allele,Stutter and Noise under different thresholds was calculated and difference comparison was performed.The results showed that the detection rate of true alleles relatively increased with the analysis threshold being stricter.（6）Analysis of ACR:The analysis threshold was defined as mentioned above.The results show that,the distribution of average ACR at each locus would be different as the analysis threshold differed.Under the 5 reads analysis threshold,the locus with the highest ACR was TPOX（0.82）,and the lowest was D20S470（0.48）.（7）Sample consistency:Detected alleles were named according to the naming guidelines of the International Society of Forensic Genetics（ISFG）and the core sequence repetition correction algorithm proposed in the relevant literature.The data of NGS-STR and CE-STR were compared to each other.It was found that the two method were consistent at 3305 loci among the 3066ones of 73 samples（98.36%）,if the analysis threshold was 5 reads and only length polymorphism was considered.And within the 3305 loci,a total of 349types of alleles was found with CE-STR method,such number would increase to 501 by NGS-STR if sequence polymorphism was taken into consideration,and 152 types of isoalleles were detected.Isoalleles appeared at 24 loci,and the locus with the most was D13S317,which accounted for 160.00%of the total number of original alleles.2.Threshold definition and performance evaluation of NGS-STR system in the 2nd-degree kinship identification（1）Distributing of IBS score and CLRThere were 47 pairs of grandparent-grandchild pairs and 87uncle-nephew pairs determined in 10 pedigrees,and a total of 83 samples were involved in these pairs.The accurate typing results of 10 samples could not be obtained owing to the poor results of sequence,so the follow-up analysis was based on the remaining 80 samples within 115pairs of 2nd-degree relatives.IBS and CLR of these pairs and randomly selected unrelated pairs of equivalent amount.There would be a large overlap between the IBS distribution in the two groups,no matter which method,CE or NGS,was used.In contrast,such overlap would be significantly reduced if CLR was applied,indicating that the identification ability of CLR was better than that of IBS.（2）Threshold to confirm 2nd-degree relationshipsTwo sets of thresholds were set applying the distribution results of Log₁₀（CLR）values for the confirmation of 2nd-degree relationships.The results of diagnostic test show that if Log₁₀（CLR）≥2 was used as the confirming threshold,the false positive rate and sensitivity of 42STRs-CE were 0.00%and 72.17%,respectively,and those of 42STRs-NGS were 0.00%and 81.74%,respectively.The false positive rates of the two systems would remain unchanged when the above threshold was 1,while the sensitivity of42STRs-CE and 42STRs-NGS would increase to 90.43%and 90.43%,respectively.Both of the two systems would have a false positive rate of0.00%under the lower threshold,and the sensitivity decreases with the increase of it.（3）Threshold to exclude 2nd-degree relationshipsIf Log₁₀（CLR）≤-1 was used as the threshold to exclude the 2nd-degree kinship,the specificity and the false negative rate of 42STRs-CE was 84.35%and 2.61%,respectively,the those of 42STRs-NGS was 89.57%and 2.61%,respectively.Such parameters would be 67.83%,0.00%,81.74%and 0.87%,respectively,if Log₁₀（CLR）≤-2 was used as the excluding threshold.With the decrease of the threshold,the specificity of the 42STRs-CE system and the42STRs-NGS system decreased.On the whole,the specificity of the42STRs-NGS system was better than that of the 42STRs-CE system when the2nd-degree kinship was excluded.In summary,the Log₁₀（CLR）value with higher sensitivity was set as the lowest threshold for confirming the 2nd-degree relationships,and that with higher specificity as the highest threshold for excluding.Under this strategy,we determined Log₁₀（CLR）≥1 and Log₁₀（CLR）≤-1 as the confirming and excluding thresholds,respectively.Under this threshold,tendentious opinion could not be given in 20（12unrelated and 8 2nd-degree relationships）of the230 pairs of samples,the power of the genotyping system was 91.30%.Within the rest 210 pairs,there were 3 misjudgments,all of which were mistakenly exclusion of true relatives,making the accuracy rate when tendentious opinion given being 98.57%.Conclusions:In this study,the NGS-STR typing system containing 42 autosomal STR and Amelogenin gene constructed in the earlier stage of the laboratory was used to detect 115 pairs of grandparents&grandchildren/uncle&nephew samples,and the forensic application value of the system in the identification of 2nd-degree kinship relationships was evaluated.The results showed that the NGS-STR system has good stability in our laboratory,good consistency with the CE results while detecting more isoallele at the same time,which improved the efficiency of the system.The classification system is used to the determination of 2nd-degree kinship,the sensitivity,specificity,false positive rate and false negative rate were 90.43%,89.57%,0.00%and2.61%,respectively,when Log₁₀（CLR）≥1 and Log₁₀（CLR）≤-1 were set as the confirming and excluding thresholds of 2nd-degree relationships,respectively.The sensitivity and specificity are both 89.52%,the false positive is 0%,and the false negativity is 2.54%;the power of the genotyping system is91.30%,and the accuracy rate is 98.57%when the preference opinion is obtained.Such results showed that the system could perform well when distinguishing grandparents&grandchildren and/or uncles&nephews from unrelated individuals,providing new basic data and technical solutions for the2nd-degree kinship identification.