Re-Analysis of Next Generation Sequencing Gene Panel Data By Normalized Coverage Values Reveals Previously Undetected Copy Number Variations in Inherited Bone Marrow Failure Syndromes

Background: Inherited bone marrow failure syndromes (IBMFSs) comprise a group of genetic disorders affecting blood cell production in bone marrow, resulting in single or multilineage cytopenias. Patients with IBMFSs often have physical malformations and are at risk of developing leukemia and solid tumors. With advances in sequencing technologies, causal nucleotide-level mutations can readily be identified in many IBMFSs. In about half of IBMFS cases, however, a molecular mutation is not detected. Copy number variation (CNVs) have been reported to cause IBMFSs. Unfortunately, genome-wide methods to identify CNVs have major limitations as they may miss small lesions (e.g. by oligonucleotide or CGH arrays) or may have low sensitivity due to low read depths (e.g. by whole genome sequencing).

Purpose: This study's purpose was to determine, after standard analysis of a next generation sequencing gene panel for causal mutations, whether re-analysis by normalized coverage value could identified CNVs and to characterize the genetic lesions.

Methods & Materials: Patients with IBMFSs were enrolled in The Canadian Inherited Marrow Failure Registry (CIMFR). Genomic DNA was analyzed by next-generation panel gene panel of over 72 IBMFS-associated genes. After analysis for nucleotide level mutations, the NextGene software was used to detect copy number variation using nucleotide read depth in test samples versus normalized control coverage values. Monoalleic copy number deletions were indicated by ratio of <0.33;biallelic deletion was considered when the value was close to zero. PCR or Affymetrix SNP6.0 array were used to validate deletions.

Results: Of the 258 patients who tested by the IBMFS next-generation gene panel, 93 (36%) were found to have causal nucleotide level mutations. 168 patients without identified mutations underwent analysis for copy number variation, of which 10 (6%) were found to have deletions. Genes most commonly exhibiting monoallelic deletions included RPS19, RPL11 and RPL5 (5 Diamond-Blackfan anemia patients). One patients with MDS was found to have monoalleic deletion of GATA2 . Moreover, a bialleic deletion of exon 1-5 in FANCA gene was detected in one Fanconi anemia patient. Lastly, we also identified monoalleic deletions of RBM8A in two thrombocytopenia with absent radii patients and of PARN in one unclassified pancytopenia patient. The last three patients were found to also have a nucleotide level mutation on the other allele by the NGS panel (compound heterozygousity). All the above deletions were validated by another method (PCR or Affymetrix SNP6.0 array).

Conclusions: CNVs can be detected in some IBMFSs by next-generation sequencing (NGS) panels by careful analysis of normalized coverage values. Such analysis should become standard practice before usage of other mutation detection strategies such as CNV microarray or whole exome/genome sequencing, since the later techniques may not detect certain CNVs that can be detected by gene panels.