Comparison of in silico strategies to prioritize rare genomic variants impacting RNA splicing for the diagnosis of genomic disorders

Rowlands, Charlie F.; Thomas, Huw B.; Lord, Jenny; Wai, Htoo A.; Arno, Gavin; Beaman, Glenda M.; Sergouniotis, Panagiotis I.; Gomes-Silva, Beatriz; Campbell, Christopher; Gossan, Nicole; Hardcastle, Claire; Webb, Kevin; O’Callaghan, Christopher; Hirst, Robert A.; Ramsden, Simon; Jones, Elizabeth A.; Clayton‐Smith, Jill; Webster, Andrew R.; Ambrose, J. C.; Arumugam, Prabhu; Bevers, R.; Bleda, Marta; Boardman-Pretty, F.; Boustred, C. R.; Brittain, Helen; Caulfield, Mark J.; Chan, G. C.; Fowler, Tom; Giess, Adam; Hamblin, Angela; Henderson, Shirley; Hubbard, Tim; Jackson, R.; Jones, Louise; Kasperavičiūtė, Dalia; Kayikci, Melis; Kousathanas, Athanasios; Lahnstein, L.; Leigh, Sarah; Leong, I. U. S.; Lopez, Fabrice; Maleady‐Crowe, F.; McEntagart, Meriel; Minneci, Federico; Moutsianas, Loukas; Mueller, Michael; Murugaesu, Nirupa; Need, Anna C.; O′Donovan, Peter; Odhams, Christopher A.; Patch, Christine; Perez-Gil, D.; Pereira, Mariana Buongermino; Pullinger, J.; Rahim, T.; Rendon, Augusto; Rogers, T.; Savage, K.; Sawant, K.; Scott, Richard; Siddiq, Afshan; Sieghart, A.; Smith, Samuel C.; Sosinsky, Alona; Stuckey, A.; Tanguy, M.; Tavares, Ana Lisa Taylor; Thomas, Ellen; Thompson, S. R.; Tucci, Arianna; Welland, M. J.; Williams, Eleanor; Witkowsa, K.; Wood, S. M.; Douglas, Andrew G.L.; O’Keefe, Raymond T.; Baralle, Diana; Black, Graeme; Ellingford, Jamie M

doi:10.1038/s41598-021-99747-2

Cited by 44 publications

(26 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Pathogenic variants, both protein-coding and intronic, that lie outside canonical splice sites may nonetheless act to disrupt pre-mRNA splicing through a diverse series of mechanisms ( Figure S1 ). 1 , 2 , 3 Effective identification of pathogenic splice-impacting variants remains challenging and is limited by the omission of intronic regions in targeted sequencing approaches, 4 , 5 discordance between in silico variant prioritization tools, 6 and the lack of availability of the appropriate tissue from which to survey RNA for splicing disruption. 7 , 8 …”

Section: Introductionmentioning

confidence: 99%

MRSD: A quantitative approach for assessing suitability of RNA-seq in the investigation of mis-splicing in Mendelian disease

Rowlands

Taylor

Rice

et al. 2022

The American Journal of Human Genetics

Self Cite

View full text Add to dashboard Cite

Variable levels of gene expression between tissues complicates the use of RNA sequencing of patient biosamples to delineate the impact of genomic variants. Here, we describe a gene-and tissue-specific metric to inform the feasibility of RNA sequencing. This overcomes limitations of using expression values alone as a metric to predict RNA-sequencing utility. We have derived a metric, minimum required sequencing depth (MRSD), that estimates the depth of sequencing required from RNA sequencing to achieve user-specified sequencing coverage of a gene, transcript, or group of genes. We applied MRSD across four human biosamples: whole blood, lymphoblastoid cell lines (LCLs), skeletal muscle, and cultured fibroblasts. MRSD has high precision (90.1%-98.2%) and overcomes transcript region-specific sequencing biases. Applying MRSD scoring to established disease gene panels shows that fibroblasts, of these four biosamples, are the optimum source of RNA for 63.1% of gene panels. Using this approach, up to 67.8% of the variants of uncertain significance in ClinVar that are predicted to impact splicing could be assayed by RNA sequencing in at least one of the biosamples. We demonstrate the utility and benefits of MRSD as a metric to inform functional assessment of splicing aberrations, in particular in the context of Mendelian genetic disorders to improve diagnostic yield.

show abstract

Section: Introductionmentioning

confidence: 99%

MRSD: A quantitative approach for assessing suitability of RNA-seq in the investigation of mis-splicing in Mendelian disease

Rowlands

Taylor

Rice

et al. 2022

The American Journal of Human Genetics

Self Cite

View full text Add to dashboard Cite

show abstract

“…On this data, CI-SpliceAI had the greatest accuracy of all tools tested on both the binary task and when predicting the exact variant effect. SpliceAI has performed favourably in many comparisons since its release in 2019 [ 16 , 17 , 20 , 24 , 29 ], and it was suggested that the simultaneous prediction of thousands of nucleotides around a variant is the key advantage for its success: the big window sizes might allow SpliceAI to recognise pairs of acceptors and donors and other co-dependent features not only near splice sites, but also deep within the exon or intron. Feature maps within a convolutional neural network that recognise patterns in data are applied as a sliding window, allowing splicing factors such as binding sites or the branch point to be recognised independent of their offset to a splice site, and variants within their motifs are considered in the classification.…”

Section: Discussionmentioning

confidence: 99%

“…Many splice prediction tools exist, but there is little consensus on which the best tools are, what thresholds should be used to detect significant disruptions, and how these applications should be applied in clinical diagnostics. Recent applications of machine learning to splice site prediction show great promise, and have been found to be more accurate than older but still widely used methods such as MaxEntScan (MES) [ 15 – 17 ]. MES models the likelihood of a splice site given 9 or 23 bases using a maximum entropy model; MMSplice [ 18 ] uses neural networks to predict splice sites given 18 or 53 nucleotides; SQUIRLS [ 19 ] uses carefully engineered features from around the splice sites to classify using decision trees; and SpliceAI [ 20 ] uses five deep convolutional neural networks to predict splice sites based on 10,000 nucleotides of context.…”

Section: Introductionmentioning

confidence: 99%

CI-SpliceAI—Improving machine learning predictions of disease causing splicing variants using curated alternative splice sites

et al. 2022

Self Cite

View full text Add to dashboard Cite

Background It is estimated that up to 50% of all disease causing variants disrupt splicing. Due to its complexity, our ability to predict which variants disrupt splicing is limited, meaning missed diagnoses for patients. The emergence of machine learning for targeted medicine holds great potential to improve prediction of splice disrupting variants. The recently published SpliceAI algorithm utilises deep neural networks and has been reported to have a greater accuracy than other commonly used methods. Methods and findings The original SpliceAI was trained on splice sites included in primary isoforms combined with novel junctions observed in GTEx data, which might introduce noise and de-correlate the machine learning input with its output. Limiting the data to only validated and manual annotated primary and alternatively spliced GENCODE sites in training may improve predictive abilities. All of these gene isoforms were collapsed (aggregated into one pseudo-isoform) and the SpliceAI architecture was retrained (CI-SpliceAI). Predictive performance on a newly curated dataset of 1,316 functionally validated variants from the literature was compared with the original SpliceAI, alongside MMSplice, MaxEntScan, and SQUIRLS. Both SpliceAI algorithms outperformed the other methods, with the original SpliceAI achieving an accuracy of ∼91%, and CI-SpliceAI showing an improvement at ∼92% overall. Predictive accuracy increased in the majority of curated variants. Conclusions We show that including only manually annotated alternatively spliced sites in training data improves prediction of clinically relevant variants, and highlight avenues for further performance improvements.

show abstract

“…Although SpliceAI is regarded as the best single method to predict aberrant splicing, the combined use of different tools can further increase accuracy. 27 , 28 It is hypothesized that the c.383-1368A>G variant creates a novel binding site for the splicing enhancer, protein SRSF1 (SF2/ASF), leading to the inclusion of a pseudo-exon containing a premature stop codon. Although a minigene splicing assay with sequence analysis of the mutated transcript exactly matched aberrant splicing as predicted by combining deep learning splice prediction tools, 22 , 23 additional families would further support and validate our results.…”

Section: Discussionmentioning

confidence: 99%

Noncanonical Splice Site and Deep Intronic FRMD7 Variants Activate Cryptic Exons in X-linked Infantile Nystagmus

Lee

Jeong

Won

et al. 2022

Trans. Vis. Sci. Tech.

View full text Add to dashboard Cite

Purpose We aim to report noncoding pathogenic variants in patients with FRMD7 -related infantile nystagmus (FIN). Methods Genome sequencing ( n = 2 families) and reanalysis of targeted panel next generation sequencing ( n = 2 families) was performed in genetically unsolved cases of suspected FIN. Previous sequence analysis showed no pathogenic coding variants in genes associated with infantile nystagmus. SpliceAI, SpliceRover, and Alamut consensus programs were used to annotate noncoding variants. Minigene splicing assay was performed to confirm aberrant splicing. In silico analysis of exonic splicing enhancer and silencer was also performed. Results FRMD7 intronic variants were identified based on genome sequencing and targeted next-generation sequencing analysis. These included c.285-12A>G (pedigree 1), c.284+63T>A (pedigrees 2 and 3), and c. 383-1368A>G (pedigree 4). All variants were absent in gnomAD, and the both c.285-12A>G and c.284+63T>A variants were predicted to enhance new splicing acceptor gains with SpliceAI, SpliceRover, and Alamut consensus approaches. However, the c.383-1368 A>G variant only had a significant impact score on the SpliceRover program. The c.383-1368A>G variant was predicted to promote pseudoexon inclusion by binding of exonic splicing enhancer. Aberrant exonizations were validated through minigene constructs, and all variants were segregated in the families. Conclusions Deep learning–based annotation of noncoding variants facilitates the discovery of hidden genetic variations in patients with FIN. This study provides evidence of effectiveness of combined deep learning–based splicing tools to identify hidden pathogenic variants in previously unsolved patients with infantile nystagmus. Translational Relevance These results demonstrate robust analysis using two deep learning splicing predictions and in vitro functional study can lead to finding hidden genetic variations in unsolved patients.

show abstract

Comparison of in silico strategies to prioritize rare genomic variants impacting RNA splicing for the diagnosis of genomic disorders

Cited by 44 publications

References 52 publications

MRSD: A quantitative approach for assessing suitability of RNA-seq in the investigation of mis-splicing in Mendelian disease

MRSD: A quantitative approach for assessing suitability of RNA-seq in the investigation of mis-splicing in Mendelian disease

CI-SpliceAI—Improving machine learning predictions of disease causing splicing variants using curated alternative splice sites

Noncanonical Splice Site and Deep Intronic FRMD7 Variants Activate Cryptic Exons in X-linked Infantile Nystagmus

Contact Info

Product

Resources

About