Significance of Near-Canonical Conserved Splice Region Nucleotides
Clinical Genetics and Therapeutics
-
Primary Categories:
- Clinical Genetics
-
Secondary Categories:
- Clinical Genetics
Introduction:
Introduction: While sequence variants disrupting the canonical splice donor and canonical splice acceptor sites are commonly treated as presumptive loss-of-function (in accordance with the ACMG/AMP PVS1 guidelines), high conservation is also found at the nearby +5 (77% G) and -3 (93% C or T) positions in the splice donor and splice acceptor regions, respectively. Based on this apparent functionally-driven nucleotide-level conservation, we performed a meta-analysis of published data to determine the extent to which variants affecting these positions disrupted canonical splicing activity, and whether partial weighting of PVS1 might be applied. Subsequently, we incorporated a post hoc analysis of internally curated variants at the +5 and -3 positions to validate in a less-selected set of variants.
Methods:
Methods: HGMD data were mined for all nucleotide substitutions involving the +5 nucleotide position and the -3 nucleotide position. Literature in the generated list was assessed for experimental evidence of splicing effects and catalogued according to type of variant and experimental outcome. Given the selection for a combination of C (64%) and T (29%) at the -3 position of the acceptor, we also examined whether C>T and T>C transitions were more functionally tolerated based on the published experimental data, and therefore potentially less likely to cause a phenotype. Internally evaluated variants from Quest Diagnostics panels were also taken into consideration as a confirmatory post hoc evaluation, as they were potentially less likely to have undergone the ascertainment bias of being publishable splice disruptions that could reasonably be expected in HGMD sources.
Results:
Results: We identified 285 nucleotide substitutions at the +5 position variants with published functional evidence. Only 12 (4.2%, 95% CI: 2.4 – 7.2%) had reports of inconclusive or insignificant effects on splicing. We also found 116 nucleotide substitutions at the -3 position variants with published functional evidence. Of these, 9 had reports of inconclusive or insignificant effects on splicing. Specifically, 4 of the 9 nucleotide substitutions explicitly found to not significantly affect splicing (3 C>T and 1 T>C) were C<->T transitions at the -3 nucleotide, which is to be expected based on the evolutionary conservation. It is also notable that even after taking into account the general publication skew towards positive findings, nearly half of -3C<->T variants in HGMD with experimental data showed no meaningful splicing disruption.
When evaluating internal variant data, we found that 70.4% (95% CI: 59.7 – 79.2%) of classified variants involving a +5G>N substitution reached pathogenic or likely pathogenic, but that 0% (95% CI: 0 – 7.3%) did when the +5 reference nucleotide was A/C/T. In internal substitutions at the -3 position, 6.7% (95% CI: 18.5 – 21.3%) of -3C/T>A and 25.6% (95% CI: 14.9 – 40.2%) of -3C/T>G variants reached pathogenic or likely pathogenic, while 0% (95% CI: 0 – 3%)of internal-3C<->T variants reached pathogenic or likely pathogenic, suggesting that -3C<->T changes are generally well-tolerated in vivo.
Conclusion:
Conclusions: Based on literature and internal post-hoc analysis of classified variants, we suggest consideration of partial PVS1 weight application to specifically +5G>N variants, but not +5A/C/T>N variants. With additional published experimental data incorporated in the future, in particular general population transcript data not under reporting selection for affected individuals, there may soon be sufficient evidence to support additional weighting of the BP4 criterion for C<->T changes at the -3 position of the acceptor region, as well as for A/C/T>N changes at the +5 position of the donor region.
Introduction: While sequence variants disrupting the canonical splice donor and canonical splice acceptor sites are commonly treated as presumptive loss-of-function (in accordance with the ACMG/AMP PVS1 guidelines), high conservation is also found at the nearby +5 (77% G) and -3 (93% C or T) positions in the splice donor and splice acceptor regions, respectively. Based on this apparent functionally-driven nucleotide-level conservation, we performed a meta-analysis of published data to determine the extent to which variants affecting these positions disrupted canonical splicing activity, and whether partial weighting of PVS1 might be applied. Subsequently, we incorporated a post hoc analysis of internally curated variants at the +5 and -3 positions to validate in a less-selected set of variants.
Methods:
Methods: HGMD data were mined for all nucleotide substitutions involving the +5 nucleotide position and the -3 nucleotide position. Literature in the generated list was assessed for experimental evidence of splicing effects and catalogued according to type of variant and experimental outcome. Given the selection for a combination of C (64%) and T (29%) at the -3 position of the acceptor, we also examined whether C>T and T>C transitions were more functionally tolerated based on the published experimental data, and therefore potentially less likely to cause a phenotype. Internally evaluated variants from Quest Diagnostics panels were also taken into consideration as a confirmatory post hoc evaluation, as they were potentially less likely to have undergone the ascertainment bias of being publishable splice disruptions that could reasonably be expected in HGMD sources.
Results:
Results: We identified 285 nucleotide substitutions at the +5 position variants with published functional evidence. Only 12 (4.2%, 95% CI: 2.4 – 7.2%) had reports of inconclusive or insignificant effects on splicing. We also found 116 nucleotide substitutions at the -3 position variants with published functional evidence. Of these, 9 had reports of inconclusive or insignificant effects on splicing. Specifically, 4 of the 9 nucleotide substitutions explicitly found to not significantly affect splicing (3 C>T and 1 T>C) were C<->T transitions at the -3 nucleotide, which is to be expected based on the evolutionary conservation. It is also notable that even after taking into account the general publication skew towards positive findings, nearly half of -3C<->T variants in HGMD with experimental data showed no meaningful splicing disruption.
When evaluating internal variant data, we found that 70.4% (95% CI: 59.7 – 79.2%) of classified variants involving a +5G>N substitution reached pathogenic or likely pathogenic, but that 0% (95% CI: 0 – 7.3%) did when the +5 reference nucleotide was A/C/T. In internal substitutions at the -3 position, 6.7% (95% CI: 18.5 – 21.3%) of -3C/T>A and 25.6% (95% CI: 14.9 – 40.2%) of -3C/T>G variants reached pathogenic or likely pathogenic, while 0% (95% CI: 0 – 3%)of internal-3C<->T variants reached pathogenic or likely pathogenic, suggesting that -3C<->T changes are generally well-tolerated in vivo.
Conclusion:
Conclusions: Based on literature and internal post-hoc analysis of classified variants, we suggest consideration of partial PVS1 weight application to specifically +5G>N variants, but not +5A/C/T>N variants. With additional published experimental data incorporated in the future, in particular general population transcript data not under reporting selection for affected individuals, there may soon be sufficient evidence to support additional weighting of the BP4 criterion for C<->T changes at the -3 position of the acceptor region, as well as for A/C/T>N changes at the +5 position of the donor region.