A Systematic Method for Detecting Abnormal mRNA Splicing and Assessing Its Clinical Impact in Individuals Undergoing Genetic Testing for Hereditary Cancer Syndromes

Abstract

Nearly 14% of disease-causing germline variants result from the disruption of mRNA splicing. Most (67%) DNA variants predicted in silico to disrupt splicing are classified as variants of uncertain significance. An analytic workflow—splice effect event resolver (SPEER)—was developed and validated to use mRNA sequencing to reveal significant deviations in splicing, pinpoint the DNA variants potentially involved, and measure the deleterious effects of the altered splicing on mRNA transcripts, providing evidence for assessing the pathogenicity of the variant. SPEER was used to analyze leukocyte RNA encoding 63 hereditary cancer syndrome–related genes in 20,317 patients. Among 3563 patients (17.5%) with at least one DNA variant predicted to affect splicing, 971 (4.8%) had altered splicing with a deleterious effect on the transcript, and 40 had altered splicing due to a DNA variant located outside of the reportable range of the test. Integrating SPEER results into the interpretation of variants allowed variants of uncertain significance to be reclassified as pathogenic or likely pathogenic in 0.4%, and as benign or likely benign in 5.9%, of the 20,317 patients. SPEER-based evidence was associated with a significantly greater effect on classifications of pathogenic or likely pathogenic and benign or likely benign in nonwhite versus non-Hispanic white patients, illustrating that evidence derived from mRNA splicing analysis may help to reduce ethnic/ancestral disparities in genetic testing. Nearly 14% of disease-causing germline variants result from the disruption of mRNA splicing. Most (67%) DNA variants predicted in silico to disrupt splicing are classified as variants of uncertain significance. An analytic workflow—splice effect event resolver (SPEER)—was developed and validated to use mRNA sequencing to reveal significant deviations in splicing, pinpoint the DNA variants potentially involved, and measure the deleterious effects of the altered splicing on mRNA transcripts, providing evidence for assessing the pathogenicity of the variant. SPEER was used to analyze leukocyte RNA encoding 63 hereditary cancer syndrome–related genes in 20,317 patients. Among 3563 patients (17.5%) with at least one DNA variant predicted to affect splicing, 971 (4.8%) had altered splicing with a deleterious effect on the transcript, and 40 had altered splicing due to a DNA variant located outside of the reportable range of the test. Integrating SPEER results into the interpretation of variants allowed variants of uncertain significance to be reclassified as pathogenic or likely pathogenic in 0.4%, and as benign or likely benign in 5.9%, of the 20,317 patients. SPEER-based evidence was associated with a significantly greater effect on classifications of pathogenic or likely pathogenic and benign or likely benign in nonwhite versus non-Hispanic white patients, illustrating that evidence derived from mRNA splicing analysis may help to reduce ethnic/ancestral disparities in genetic testing. The disruption of normal mRNA splicing is a common cause of genetic disease. Approximately 14% of germline pathogenic or likely pathogenic (P/LP) variants cause disease through mis-splicing,1Truty R. Ouyang K. Rojahn S. Garcia S. Colavin A. Hamlington B. Freivogel M. Nussbaum R.L. Nykamp K. Aradhya S. Spectrum of splicing variants in disease genes and the ability of RNA analysis to reduce uncertainty in clinical interpretation.Am J Hum Genet. 2021; 108: 696-708Abstract Full Text Full Text PDF PubMed Scopus (27) Google Scholar with the vast majority (>90%) of these P/LP splicing variants disrupting the canonical donor and acceptor splice sites (first 1 to 2 bp flanking an exon). In contrast, many DNA variants other than in the canonical donor and acceptor dinucleotide splice sites are predicted by commonly used algorithms to affect splicing, but most (69%) of these potential splicing variants (PSpVs) are classified as variants of uncertain significance (VUSs),1Truty R. Ouyang K. Rojahn S. Garcia S. Colavin A. Hamlington B. Freivogel M. Nussbaum R.L. Nykamp K. Aradhya S. Spectrum of splicing variants in disease genes and the ability of RNA analysis to reduce uncertainty in clinical interpretation.Am J Hum Genet. 2021; 108: 696-708Abstract Full Text Full Text PDF PubMed Scopus (27) Google Scholar demonstrating the challenge of classifying non–canonical donor and acceptor splice site PSpVs in the absence of confirmatory RNA testing. Knowing which of these PSpVs have a clinically significant effect on splicing could inform the interpretation of variants, resulting in fewer VUSs and increased actionability in genetic testing.2Smirnov D. Schlieben L.D. Peymani F. Berutti R. Prokisch H. Guidelines for clinical interpretation of variant pathogenicity using RNA phenotypes.Hum Mutat. 2022; 43: 1056-1070Crossref PubMed Scopus (4) Google Scholar Here, a novel method is described: the splice effect event resolver (SPEER), which uses RNA sequencing to determine whether variants found during DNA sequencing cause aberrant mRNA splicing and have a deleterious impact on the structure and function of transcripts. SPEER accomplishes this task in three steps: i) determining whether splicing is abnormal compared to that in controls, ii) quantifying the reduction in normal splicing at a crucial splice junction, and iii) examining the consequences of abnormal-splicing events on mRNA structure and function. SPEER begins with short-read sequencing of cDNA and several overlapping RNA-based analytical methods,3Li Y.I. Knowles D.A. Humphrey J. Barbeira A.N. Dickinson S.P. Im H.K. Pritchard J.K. Annotation-free quantification of RNA splicing using LeafCutter.Nat Genet. 2018; 50: 151-158Crossref PubMed Scopus (283) Google Scholar, 4Frésard L. Smail C. Ferraro N.M. Teran N.A. Li X. Smith K.S. Bonner D. Kernohan K.D. Marwaha S. Zappala Z. Balliu B. Davis J.R. Liu B. Prybol C.J. Kohler J.N. Zastrow D.B. Reuter C.M. Fisk D.G. Grove M.E. Davidson J.M. Hartley T. Joshi R. Strober B.J. Utiramerur S. Lind L. Ingelsson E. Battle A. Bejerano G. Bernstein J.A. Ashley E.A. Boycott K.M. Merker J.D. Wheeler M.T. Montgomery S.B. Identification of rare-disease genes using blood transcriptome sequencing and large control cohorts.Nat Med. 2019; 25: 911-919Crossref PubMed Scopus (157) Google Scholar, 5Mertes C. Scheller I.F. Yépez V.A. Çelik M.H. Liang Y. Kremer L.S. Gusic M. Prokisch H. Gagneur J. Detection of aberrant splicing events in RNA-seq data using FRASER.Nat Commun. 2021; 12: 529Crossref PubMed Scopus (49) Google Scholar, 6Jenkinson G. Li Y.I. Basu S. Cousin M.A. Oliver G.R. Klee E.W. LeafCutterMD: an algorithm for outlier splicing detection in rare diseases.Bioinformatics. 2020; 36: 4609-4615Crossref PubMed Scopus (22) Google Scholar, 7Vaquero-Garcia J. Barrera A. Gazzara M.R. González-Vallinas J. Lahens N.F. Hogenesch J.B. Lynch K.W. Barash Y. A new view of transcriptome complexity and regulation through the lens of local splicing variations.Elife. 2016; 5: e11752Crossref PubMed Scopus (221) Google Scholar to analyze exon–exon junctions and to determine whether the alterations observed in the mRNA-splicing pattern of a gene containing a PSpV are statistically significant relative to those in a panel of controls. In this way, SPEER distinguishes variant-specific alterations in splicing from the naturally occurring alternative splicing that generates protein diversity and function specificity across tissue types.8Sibley C.R. Blazquez L. Ule J. Lessons from non-canonical splicing.Nat Rev Genet. 2016; 17: 407-421Crossref PubMed Scopus (160) Google Scholar, 9Baralle F.E. Giudice J. Alternative splicing as a regulator of development and tissue identity.Nat Rev Mol Cell PubMed Scopus Google Scholar, J. B.J. Alternative splicing and 2019; Full Text Full Text PDF PubMed Scopus Google Scholar a for quantifying the to which normal splicing is in a compared to controls. SPEER overlapping into one or of abnormal-splicing which be for and potential mRNA the of the deleterious impact of a DNA variant on structure and the evidence by SPEER be used for the interpretation of variants a as K. M. M. Garcia J. B. Y. J. M. S. a of the variant Med. Full Text Full Text PDF PubMed Scopus Google Scholar that many other of evidence to the pathogenicity of a variant. on the of the of SPEER to events and in a of patients for hereditary cancer syndrome–related genes to which PSpVs during DNA sequencing significantly altered splicing and had a deleterious effect on structure and function. SPEER abnormal-splicing in a of patients with DNA variants located outside of the reportable range of a sequencing DNA test. SPEER-based evidence was used in for the interpretation of variants, a significant of PSpVs as P/LP or as benign or likely benign most SPEER allowed for an greater of of PSpVs patients as nonwhite compared to those as non-Hispanic In 1 of this a of data to be used for SPEER was had genetic for hereditary cancer at and had a germline DNA variant that was or predicted to splicing and have a but had a or of a hereditary cancer as in was a by the a from and blood to be used in genetic and was to of aberrant splicing splicing in nonwhite versus white by to into potential in splicing of germline of as in in The M. T. X. K. J.A. alternative RNA splicing and events from and white patients in The 2021; PubMed Scopus Google Scholar In a of 20,317 patients to for hereditary gene DNA and RNA sequencing and SPEER was used to analyze the leukocyte RNA of 63 hereditary genes in these patients. The use of and data was by an A of hereditary genes for RNA analysis of these from the i) in leukocyte are at a to a statistically significant of a splicing ii) are genes in which a of effect of abnormal a of an increased for or iii) have a variant associated with an increased for of gene was as R. Ouyang K. Rojahn S. Garcia S. Colavin A. Hamlington B. Freivogel M. Nussbaum R.L. Nykamp K. Aradhya S. Spectrum of splicing variants in disease genes and the ability of RNA analysis to reduce uncertainty in clinical interpretation.Am J Hum Genet. 2021; 108: 696-708Abstract Full Text Full Text PDF PubMed Scopus (27) Google R. J. M. E. Nussbaum R.L. Aradhya S. and of in disease Med. 2019; Full Text Full Text PDF PubMed Scopus Google M.A. L. Y. M. J.M. of a sequencing panel for hereditary cancer PubMed Scopus Google Scholar using DNA to bp of flanking and of clinical as the reportable to a of and a of at the reportable Full gene and the interpretation of variants at as K. M. M. Garcia J. B. Y. J. M. S. a of the variant Med. Full Text Full Text PDF PubMed Scopus Google M.A. L. Y. M. J.M. of a sequencing panel for hereditary cancer PubMed Scopus Google Scholar PSpVs predicted using G. of with to RNA splicing PubMed Scopus Google Scholar RNA splice junctions of of and in gene PubMed Scopus Google Scholar and the in the PSpVs predicted using K. S. Knowles D. Li Y.I. J.A. J. E. H. A. S. splicing from with 2019; Full Text Full Text PDF PubMed Scopus Google Scholar RNA was from blood in RNA and containing that RNA using and by DNA was by cDNA using the RNA F. and of with DNA by and during and for 2 bp from sequencing and on the with A. Davis F. J. C. S. M. RNA-seq PubMed Scopus Google Scholar using a using T. A. sequencing in to quantification PubMed Scopus Google from the using the and in Y.I. Knowles D.A. Humphrey J. Barbeira A.N. Dickinson S.P. Im H.K. Pritchard J.K. Annotation-free quantification of RNA splicing using LeafCutter.Nat Genet. 2018; 50: 151-158Crossref PubMed Scopus (283) Google Scholar SPEER used sequencing of RNA from blood to determine across 63 genes of the significance of in a SPEER A and the of and novel splice junctions using in Y. and of RNA sequencing for PubMed Scopus Google S. K. M. Alternative splicing in RNA-seq in Hum Genet. PubMed Scopus Google Scholar as the of the of the by the of overlapping the SPEER the significance of splicing in a by to the from the panel using a in and was significantly from that of the control panel at SPEER uses these significant junctions as to other altered junctions that in to junctions into a of abnormal-splicing for observed in at least one a normal was using the In this the of that the was as the of and the of that the was the of junction, the the and measure of the increased relative to the to and as a panel of normal splicing The control of compared to the in to a using a test. was by the algorithm for DNA variants from A. M. S. A. J.M. L. A. S. M.R. and of in normal PubMed Scopus Google Scholar abnormal observed in a a reduction in the of one or canonical splice and an in of one or or splice junctions measure the reduction in normal splicing associated with a a in was as of the canonical to be by the of that in the control The was to with the of the normal or most commonly a significant reduction in was to with pathogenicity for PSpVs in genes to cause disease with an canonical junctions as for the canonical that was by splicing the that the PSpV was for the canonical with the in the control panel that a with a was the of a canonical was than that of other canonical junctions with the across canonical junctions was with a reduction in and at least one bp of a PSpV for by variant interpretation was for assessing whether the by junctions may have for the reduction in a PSpV that a acceptor splice may have in a a acceptor splice site resulting in a novel protein with a this novel protein for the of the protein is a of the The in this was to splice in an event as i) canonical by in the canonical ii) junctions with one iii) of a canonical was in the of an was in a and junctions an with a a of the classified as to result in or in the transcript, or of the junction, classified as with and with predicted effects on protein classified as variants located outside of the reportable range bp flanking of the are by SPEER an abnormal at the variant and one or abnormal junctions a of the data for a DNA variant that may have the abnormal-splicing was found in the DNA sequencing was the the abnormal-splicing a PSpV was in the the abnormal-splicing was for interpretation by the a PSpV could be the abnormal-splicing event was to be a or the result of a in splicing, and was in the clinical the of SPEER-based evidence for variant a new of evidence was the K. M. M. Garcia J. B. Y. J. M. S. a of the variant Med. Full Text Full Text PDF PubMed Scopus Google Scholar evidence was to as observed RNA effects this evidence was pathogenic or benign and by the on observed RNA effects to PSpVs observed in the of in classifications in 2 of the on the of the for observed RNA of PSpV to and to observed during 2 of the by the for observed RNA effects and whether the was from P/LP or to in these for significance using the by the of patients with and in the most commonly ethnic/ancestral to those in other ethnic/ancestral and SPEER was validated using containing DNA variants, as with variants to splicing to the with variants in the or but and with variants located of a significant abnormal-splicing event A was as significant abnormal-splicing event the panel for a canonical bp of a derived from of The was the that had as in one as a of by analysis of other splice junctions a PSpV The of of with a PSpV that a normal splicing pattern on a of RNA to the absence of a splicing The of the for and of SPEER-based of the in and in the and and SPEER a significant splicing in of DNA variants in the a specificity of and of was as and was as of was to specificity the vast majority of for splice of the with of the used to and in across of the and of the results The that the a splice at a and at the used for splicing effects and the variants found in SPEER the detection of variants in and outside of the reportable in patients from the 1 of had with a of for a hereditary cancer SPEER abnormal-splicing events associated with PSpVs with The was to as a measure of the of normal splicing at the canonical most to the PSpV compared to controls. The of was for variants classified as P/LP due to a clinically significant splice and those to be which could have to have or effects on splicing The the of was demonstrating that for PSpVs to be deleterious or from as P/LP or of evidence to be used for the for quantifying the effect of a variant on splicing of the most to that in in which a DNA that one or in a of at the flanking splice In these variants, the in was to that a of was with a of canonical from one A but statistically significant in was observed was was in of the was the of to the pathogenicity of a The was in of classified P/LP variants, and in of classified variants, for an of of and of variants with a of associated with that to result in predicted to result in or the abnormal that a reduction in normal mRNA in these may be for the protein that is from the abnormal mRNA function for for this PSpVs associated with a of and an that was to cause from the the was for P/LP is to that splicing that to have a of deleterious for the protein than and in the for this the pathogenic in was increased for altered splicing that was to cause compared with altered splicing that an mRNA with or PSpVs classified as P/LP associated with and variants as or that the of pathogenicity was likely abnormal splicing. these variants from the the of a of was that PSpVs with a of are likely to cause disease through a splicing for in the statistically significant in splicing with a of to was for is to the of a of this evidence be used as evidence for a benign In 2 of the 20,317 blood for DNA sequencing and for RNA and the 3563 patients (17.5%) had at least one DNA variant by that was a PSpV splicing, by a statistically significant event with PSpVs in a of 971 patients of PSpVs in of the 63 genes by SPEER splicing was associated with a PSpV in patients of patients with a illustrating the for the algorithms and used to the RNA a and normal and abnormal to to outside of the reportable benign or likely pathogenic or likely potential splicing variant of uncertain significance. in a new of PSpVs the and abnormal and normal benign or likely pathogenic or likely potential splicing variants of uncertain significance. in a new benign or likely pathogenic or likely potential splicing variant of uncertain significance. benign or likely pathogenic or likely potential splicing variants of uncertain significance. of the of PSpVs in to the the of P/LP variants relative to VUSs the canonical donor and acceptor splice sites In was an at and from the canonical donor splice and a at and of the canonical acceptor site that most generates a novel acceptor splice abnormal-splicing events associated with PSpVs with of these PSpVs located the reportable range of bp flanking the PSpV associated with an abnormal by SPEER was classified using the evidence on observed RNA effects the interpretation of PSpVs observed in this classified as P/LP or the PSpVs classified as VUSs abnormal splicing was with a of P/LP variants and a of variants normal splicing was with a of P/LP variants and a of variants of patients from the had a P/LP or variant that have classified as a by the interpretation for the mRNA splicing data from the the PSpVs of patients of have classified as VUSs of and the PSpVs of patients of have classified as VUSs a of these data across the of classifications or statistically greater in the that as or compared with those as non-Hispanic white due to the of mRNA splicing data and and in a new The most splicing in the was with and events the most with and events the PSpV classifications and that these the most events to Among the patients in the in 1 P/LP variants found through clinical a or of a hereditary cancer found to have a variant outside of the reportable range with a deleterious effect on splicing. In the of 20,317 patients for hereditary gene 40 variants outside of the reportable range that could have associated with abnormal splicing these variants, in patients, classified as a of P/LP variants in the reportable range of the in the of 20,317 An 2 variants, in 2 patients classified as and of the variants, in patients classified as VUSs The an and to and quantifying splicing alterations in the blood leukocyte of 63 hereditary cancer syndrome–related genes in patients genetic for these The that RNA analysis to DNA sequencing of hereditary cancer syndrome–related genes the interpretation of DNA variants or predicted to splicing. In of the variants in the that splicing, located outside of the reportable range of bp flanking the of the panel and by RNA analysis to a of DNA variants that have classified as VUSs the RNA analysis in of patients in a for hereditary of had variants classified as had variants classified as the these of patients with due to RNA analysis the 20,317 patients in the was greater than that in a of patients DNA and RNA for hereditary cancer versus C. A. L. H. D. D. B. H. H. J. K. R. J. K. M. C. A. R. and splicing in a of patients for hereditary Med. 2022; PubMed Scopus (4) Google Scholar was likely the result of the of genes by RNA in the compared to the versus the of classifications on RNA analysis was significantly greater in the that as or than in the non-Hispanic white The impact of RNA analysis on in these may have from RNA analysis providing evidence that was by of patients of in variant or in the RNA analysis in the interpretation of DNA variants predicted to impact splicing may reduce the in ethnic/ancestral and reduce disparities in genetic testing. results the and in the of mRNA splicing analysis into the used for variants in genetic testing. Alternative splicing is and be to pinpoint splicing alterations associated with a hereditary disease. of in the had at least one statistically significant splicing at the compared to the panel of control was was to determine whether a PSpV altered normal splicing. on splicing alterations with a PSpV bp of the altered junction, the of abnormal-splicing events was to of in abnormal-splicing events to be associated with DNA variants in the of with mRNA splicing the impact and clinical of DNA variants associated with a significant splicing are to resulting in In DNA variants found outside of the reportable range of the with RNA analysis observed in a and the evidence for a as data from a of 20,317 patients in the had variants outside of the bp reportable altered splicing, and have classified as of P/LP variants that at bp into the is with the in RNA for hereditary genes or of C. A. L. H. D. D. B. H. H. J. K. R. J. K. M. C. A. R. and splicing in a of patients for hereditary Med. 2022; PubMed Scopus (4) Google Scholar 2 of 20,317 patients had variants outside of the bp reportable range that altered splicing but that classified as the that a variant that an effect on splicing may cause disease. in a deleterious effect on splicing to pathogenicity is that abnormal splicing be A DNA variant may result in a reduction in normal mRNA splicing that from to of normal splicing from the variant with in as to as S. B. A. G. J. E. T. K. of pathogenicity for the splice variant in in a with variant 2021; M. M. D. Identification and of aberrant splicing in disease using a Mutat. 36: PubMed Scopus Google Scholar a PSpV that to a of normal splicing is likely to cause disease than is a PSpV that to a the of PSpVs using abnormal-splicing data a measure of the of a was developed to the reduction in normal mRNA splicing alterations in blood associated with a of with of normal splicing from one to cause disease and as pathogenic evidence for hereditary cancer the is is a of the in PSpVs with abnormal-splicing 14% of variants the and of P/LP variants are the a of in the of PSpVs associated with abnormal splicing for clinical is that the effects of abnormal splicing on and protein and these into a for DNA DNA and RNA the of patients with be increased a of is the other with a of the associated with a of was to P/LP and variants, as by a was a to the of a result in these clinical patients, the was in statistically significant abnormal benign in the PSpVs associated with these greater be classified as VUSs evidence splicing a this that data from mRNA splicing in the and of variants, the results be in the of the and the 63 hereditary genes in this genes that are in the RNA of gene was to have and that from those in blood a abnormal splicing observed in may have a in the tissue of the absence of a splicing in blood may have with an absence of abnormal splicing in the the SPEER algorithm was to in A of splicing alterations results in which detection by the SPEER algorithm in P/LP DNA variants to splicing in this had a of with that likely due to a of the mRNA with development of the SPEER algorithm is to and using with which the uncertainty of variants with and aberrant junctions in the variant across a DNA variants of Most of these variants with the of as DNA variants, of the impact of the DNA variant on splicing. is to a which splicing are by but with for the DNA variants, in the DNA that the and events these genes may In this RNA was to data the in the of pathogenic variants derived from RNA these a that and be whether to RNA with DNA analysis or whether RNA be by a to patients with PSpVs in the reportable range of a a of be by patients have potential splicing variants located outside of the reportable to as In the this to of DNA and RNA for hereditary with fewer a splicing variant that could be classified as pathogenic or likely in could be by RNA analysis in patients have a or uncertain but have a for hereditary cancer on In RNA analysis of from blood is a in and assessing the impact of DNA variants that have effects on mRNA splicing and that protein function and RNA analysis is for in patients in DNA variants are predicted to affect splicing, and the of variants outside of the reportable range of the test. the splicing is and are in the and in the of the of splicing. DNA and RNA sequencing with the use of the validated SPEER data to be from patients for hereditary cancer syndrome–related gene into the clinical of splicing alterations be and the interpretation of DNA variants be to a reduction in uncertainty in for hereditary with with with with

Topics & Concepts