DEVELOPMENT OF A BIOINFORMATIC PIPELINE FOR NEXT-GENERATION SEQUENCING DATA ANALYSIS: Application to circulating cell-free fetal DNA

Prenatal diagnosis (PD) of single gene disorders and aneuploidies is currently carried out through chorionic villus sampling (CVS) and amniocentesis, two invasive procedures which have an associated risk of fetal loss of 0.5-2%. The discovery, in 1997, that cell-free fetal DNA (cffDNA) circulates in maternal plasma during gestation, has offered an alternative source of fetal genetic material which can be collected by non invasive approach. In a study aimed at developing a protocol for Non Invasive Prenatal Diagnosis of β-thalassemia, we have developed a bioinformatic pipeline to infer the fetal genotype by next generation sequencing of cffDNA. The approach consists in target sequencing of a region extending 63Kb in the β-globin gene cluster, including the causative mutation (β039) and SNPs with high heterozygosity, which are used for parental haplotypes construction. In the first part of our protocol the DNAs from parents and fetus (obtained from CVS) are sequenced in order to identify informative SNPs which are then sequenced in the corresponding cffDNA sample. The pipeline firstly analyzes sequencing data from parents and constructs the parental haplotypes based on a reference panel composed by the 1000 Genomes Reference Panel merged with sequencing data of 39 Sardinian TRIOs (father, mother, child). Secondly, the fetal genotype prediction is carried out by an algorithm that infers the most likely inherited paternal and maternal alleles and, through Viterbi algorithm, reconstructs the possible haplotype blocks inherited by the fetus. Finally, the pipeline predicts the fetal HBB genotype and validates the results comparing it with the corresponding CVS sample genotype. We processed 39 cffDNA samples, however in 7 of them the pipeline could not complete the prediction because of the lack of informative sites or low fetal fraction. In 24 out of 30 samples analyzed by the pipeline (80%) the fetal genotype was correctly determinated. In the remaining 6 samples (20%) the analysis returned only the correct paternal inherited haplotype determination. In general, obtained results are encouraging and confirm that NIPD is also feasible in couples who are at risk for a monogenic disorder and share the same variant.