Sequencing-free relative haplotype dosage analysis by digital PCR for noninvasive prenatal testing of β-thalassemia: a multicenter blinded validation study

Abstract Background: β-thalassemia is among the most common autosomal recessive monogenic disorders in southern China, Southeast Asia and the Mediterranean basin, yet noninvasive prenatal testing (NIPT) for monogenic disorders remains clinically underimplemented because fetal single-nucleotide variants must be quantified against an abundant maternal cell-free DNA (cfDNA) background. Relative mutation dosage, often implemented by digital PCR (dPCR), measures pathogenic variants directly but is vulnerable to maternal background interference, whereas relative haplotype dosage (RHDO) improves robustness by aggregating dosage signals across linked loci but generally requires sequencing-based phasing, limiting accessibility in the resource-limited regions where the disease burden is highest. We hypothesized that patient-specific haplotypes could be resolved experimentally rather than computationally, allowing RHDO to be performed entirely on a benchtop dPCR platform. Methods: We developed a sequencing-free dPCR-based RHDO assay (dRHDO) for NIPT of β-thalassemia, combining multicolor melting curve analysis to identify informative single-nucleotide polymorphisms (SNPs), a Drop-Phase assay to resolve SNP–mutation linkage experimentally, and multiplex preamplification with dPCR to quantify alleles in maternal plasma cfDNA; fetal fraction was estimated internally from parental genotype-discordant SNPs. Results: In a blinded multicenter cohort of 96 β-thalassemia families, dRHDO determined fetal genotype with 97.92% (94/96) overall accuracy; all 43 samples meeting the prespecified analytical conditions (cfDNA input ≥5 ng, fetal fraction ≥5%, ≥4 maternally informative loci, ≥1 paternally informative locus) were correctly classified (100%, 43/43), and accuracy was 96.23% (51/53) among the 53 samples not fully meeting these conditions. Both discordant cases were attributable to insufficient informative-locus redundancy. The workflow is completed in 7.25 h at an estimated cost of approximately $40 per family and requires no sequencing infrastructure. Conclusions: dRHDO converts relative haplotype dosage from a sequencing-dependent computational strategy into an assay-driven measurement, enabling accurate NIPT for β-thalassemia without sequencing infrastructure.