Whole Genome Sequencing of Diamond Blackfan Anemia Syndrome Patients Detects Mutations That Alter mRNA Splicing

Abstract

Diamond Blackfan anemia syndrome (DBAS) is a rare, heritable bone marrow failure syndrome characterized by severe macrocytic anemia, congenital anomalies and predisposition to cancer, most often diagnosed during infancy. More than 98% of DBAS patients with a molecular diagnosis have mutations in a gene encoding one of the ~80 ribosomal proteins (RP) leading to haploinsufficiency. A molecular diagnosis in a patient with DBAS is critical for a definitive diagnosis, the identification of compatible related transplant donors, and developing reproductive strategies for families. Targeted sequencing of RP genes, single nucleotide polymorphism comparative genome hybridization (SNP array) to detect >30 kb deletions (Farrar et al. Blood. 2011) and exome sequencing (WES) (Ulrisch et al. Am J Hum Genet. 2018) has identified RP mutations in ~80% of patients, leaving ~20% of patients with DBAS without a molecular diagnosis.Targeted sequencing and WES focus on only coding sequences. We hypothesized that remaining 20% of DBAS mutations were in the non-coding regions of RP genes, such as promoters or introns. To test this hypothesis, we collected DNA with informed consent for whole genome sequencing (WGS) analysis from 14 patients with no molecular diagnosis after targeted sequencing, SNP array or WES. On average, we aligned ~3.2x10 7 paired end reads of 250 base pairs for each patient (~65X coverage). We focused our analysis on the sequences in and around the RP genes. To identify deletions, we used a suite of detection tools: DELLY, GRIDSS, MANTA, and LUMPY. More than 90% of deletions identified by any 2 of these tools were confirmed by PCR.We identified 5 deletions in the introns of RP genes, ranging from 11 to 467 base pairs in length, which we hypothesized disrupted splicing of the nascent RNA transcript. To test this, we created minigenes in which we replaced exon 2 of a gamma globin gene with either the WT or mutant RP exon. All wild type exons spliced normally. A 467 base pair deletion in RPL27 exon 3 was sufficient to prevent the correct splicing of that intron. Examination of the eCLIP data for RNA binding proteins revealed that spliceosome complex proteins (including SF3B1, SF3B4 and EFTUD2) and Dead-box RNA helicases bind in the deleted region. A 28 base pair deletion in exon 3 of RPL6 removes a polypyrimidine tract that is a critical part of the 3' splice junction consensus sequence, which we presume is also deleterious. The other 3 intronic deletions did not disrupt splicing.We also identified 2 causative point mutations. A point mutation 5 bases into intron 1 of the RPS26 gene changes a base in the 5' splice donor consensus sequence, which activated a cryptic splice donor in the 5' untranslated region. This aberrant splice removes the ATG initiation codon causing an untranslatable RNA. In another patient, we identified a mutation in exon 1 of the RPS27 gene, judged to be a benign amino acid change. This mutation disrupted splicing.by activating a cryptic splice donor site in the 5' untranslated region which removes the ATG initiation codon and causes a frame shift.We were referred two patients with possible duplications of the RPL35a gene. To identify duplications, we employed MinION long read single molecule sequencing. We had an average read length of ~ 6-10kb with the longest read being 1.3Mb. Overall coverage was >85X. We used minimap2 to align the reads to the reference human genome and used SNIFFLES to call the variants. One patient was the parent of DBAS-affected patient with no history of anemia. In this patient, we identified a duplication of 400 kb that included the entire RPL35a region along with genes on either side. We conclude that this duplication is not likely to cause DBA. The second patient was diagnosed with DBAS. In this patient, we identified a duplication of 4 kb including exons 1 and 2 of RPL35a We conclude that this duplication disrupts the RPL35a gene and is a likely cause of DBA.Whole genome sequencing of 15 DBAS patients identified 5 likely causative mutations in RP genes, confirming that most genetically undiagnosed cases of DBAS will involve known genes encoding RP. We conclude that the pipeline for obtaining a molecular diagnosis for DBAS from targeted sequencing, SNP array, and exome sequencing to whole genome sequencing. DisclosuresVlachos: Novartis: Membership on an entity's Board of Directors or advisory committees. Lipton: Celgene: Membership on an entity's Board of Directors or advisory committees.