Albertine Donker

General Discussion 317 9 conditions may determine the phenotype of IRIDA, as suggested by mice studies that show and increases susceptibility to the development of ID after exposure to an iron-deficient diet in TMPRSS6 haplo-insufficient mice during periods of rapid growth and increased erythropoiesis. Concerning the possible intronic variants in the as wild type considered allele in IRIDA patients in whom only a mono-allelic TMPRSS6 defect has been identified, sequencing of the whole TMPRSS6 gene including the introns may be an appropriate approach. However, the interpretation of deep intronic alterations still forms a big challenge, compared to modifications in coding regions that cause clearly interpretable changes in amino acid sequence or splicing sites. 18 Addition of RNA studies could be helpful in determining (non)-pathogenicity of variants, especially with regards to identifying aberrant splicing. However, since RNA functions as a temporary intermediate, enabling the transfer of data from the genome to the organism, it is subject to targeted degradation by various ribonucleases. This intrinsic unstable character of RNA renders its laboratory use for diagnostic purposes challenging, especially with regards to bio-specimens stored in biobanks, since commonly used banking modalities are not able to adequately preserve RNA molecules. 74 We therefore think that at the current state of knowledge sequencing of the whole TMPRSS6 gene is actually not a feasible approach. Apart from the possibility we missed defects in the as wild type considered TMPRSS6 allele, we might have overlooked pathogenic variants in other trans -acting genes that affect TMPRSS6 expression or genes that modulate hepcidin levels independently from matriptase 2. This latter suggestion is supported by the case description of Pagani et al who described a digenic form of IRIDA due to a combination of defects in both TMPRSS6 defect and in ACVR1A , encoding the BMP receptor ALK2. 75 Therefore, we propose WES sequencing of these candidate genes that are described to be involved in hepcidin regulation in mono-allelic IRIDA cases. Of note, also this approach will inevitably generate findings that are difficult to interpret. Deletions, frameshifts, nonsense mutations and changes to splice sites are very likely to be pathogenic, but deciding whether a nucleotide substitution that has not been described earlier in SNP databases (as the NCBI dbSNP, describing information on allele frequency, functional consequence and clinical significance 76 ) is pathogenic and hence represents one of the sought after defects, can be very challenging. 76,77 Co-segregation of a certain allele variant with the relevant IRIDA phenotype within