Computational Tools for Splicing Defect Prediction in Breast/Ovarian Cancer Genes: How Efficient Are They at Predicting RNA Alterations?

Abstract

In silico tools for splicing defect prediction have a key role to assess the impact of variants of uncertain significance. Our aim was to evaluate the performance of a set of commonly used splicing in silico tools comparing the predictions against RNA in vitro results. This was done for natural splice sites of clinically relevant genes in hereditary breast/ovarian cancer (HBOC) and Lynch syndrome. A study divided into two stages was used to evaluate SSF-like, MaxEntScan, NNSplice, HSF, SPANR, and dbscSNV tools. A discovery dataset of 99 variants with unequivocal results of RNA in vitro studies, located in the 10 exonic and 20 intronic nucleotides adjacent to exon–intron boundaries of BRCA1, BRCA2, MLH1, MSH2, MSH6, PMS2, ATM, BRIP1, CDH1, PALB2, PTEN, RAD51D, STK11, and TP53, was collected from four Spanish cancer genetic laboratories. The best stand-alone predictors or combinations were validated with a set of 346 variants in the same genes with clear splicing outcomes reported in the literature. Sensitivity, specificity, accuracy, negative predictive value (NPV) and Mathews Coefficient Correlation (MCC) scores were used to measure the performance. The discovery stage showed that HSF and SSF-like were the most accurate for variants at the donor and acceptor region, respectively. The further combination analysis revealed that HSF, HSF+SSF-like or HSF+SSF-like+MES achieved a high performance for predicting the disruption of donor sites, and SSF-like or a sequential combination of MES and SSF-like for predicting disruption of acceptor sites. The performance confirmation of these last results with the validation dataset, indicated that the highest sensitivity, accuracy, and NPV (99.44%, 99.44%, and 96.88, respectively) were attained with HSF+SSF-like or HSF+SSF-like+MES for donor sites and SSF-like (92.63%, 92.65%, and 84.44, respectively) for acceptor sites.We provide recommendations for combining algorithms to conduct in silico splicing analysis that achieved a high performance. The high NPV obtained allows to select the variants in which the study by in vitro RNA analysis is mandatory against those with a negligible probability of being spliceogenic. Our study also shows that the performance of each specific predictor varies depending on whether the natural splicing sites are donors or acceptors.

Opportunistic testing of BRCA1, BRCA2 and mismatch repair genes improves the yield of phenotype driven hereditary cancer gene panels

“…In a diagnostic and rapid‐turnover setting, diagnostic laboratories may need to focus on testing the most clinically useful genes. In addition, large comprehensive panels lead to a higher rate of VUS and require continuous research‐based efforts in reclassification . A balance of diagnostic and research goals are encouraged in order to progress in this field.…”

BRCA1 ‐ and BRCA2 ‐specific in silico tools for variant interpretation in the CAGI 5 ENIGMA challenge

“…To score the effect on splicing of the CAGI variants from the ENIGMA challenge, we have used the results of our recent work (Moles‐Fernández et al, ) where we identified the best combination of in silico tools for predicting splice site alterations, among those predictors available in the package Alamut Visual v2.10. More precisely, we showed that the HSF+SSF‐like combination (with Δ‐2% and Δ‐5% as thresholds, respectively) for donor sites and the SSF‐like (Δ‐5%) for acceptor sites, exhibited an optimal performance in a benchmark combining RNA in vitro testing and a dataset of variants retrieved from public databases and reported in the literature.…”

“…We focused our efforts on the set of BRCA1 and BRCA2 variants provided by the ENIGMA consortium (Spurdle et al, 2012), and we submitted four sets of predictions per protein (Table S1). These four sets correspond to different combinations of our approaches for the prediction of variants leading to affected splicing (AS; one method; Moles-Fernández et al, 2018) or affecting protein function/structure (two methods: MLR and NN). They are the following: 1 Note on terminology: We have italicized the gene symbols (BRCA1 and BRCA2) and not the protein symbols (BRCA1 and BRCA2).…”

“…To score the effect on splicing of the CAGI variants from the ENIGMA challenge, we have used the results of our recent work (Moles-Fernández et al, 2018) where we identified the best combination of in silico tools for predicting splice site alterations, among those predictors available in the package Alamut Visual v2.10.…”

“…MLR and NN refer to our two protein-specific predictors, based on a multiple linear regression model and a neural network model, respectively. AS refers to the procedure to predict variants resulting in affected splicing (Moles-Fernández et al, 2018). Abbreviation: CAGI 5, fifth Critical Assessment of Genome Interpretation estimates becomes an important factor, since improvements are expected to be small and hard to establish.…”