Editing Intron (section)

== On the accuracy of splicing ==

The spliceosome is a very complex structure containing up to one hundred proteins and five different RNAs. The substrate of the reaction is a long RNA molecule and the transesterification reactions catalyzed by the spliceosome require the bringing together of sites that may be thousands of nucleotides apart.<ref>{{cite journal | vauthors = Wan R, Bai R, Zhan X, Shi Y | date = 2020 | title = How is precursor messenger RNA spliced by the spliceosome? | journal = Annual Review of Biochemistry | volume = 89 | pages = 333–358 | doi = 10.1146/annurev-biochem-013118-111024  | pmid = 31815536 | s2cid = 209167227 }}</ref><ref>{{cite journal | vauthors = Wilkinson ME, Charenton C, Nagai K | date = 2020 | title = RNA splicing by the spliceosome | journal = Annual Review of Biochemistry | volume = 89 | pages = 359–388 |  doi = 10.1146/annurev-biochem-091719-064225| pmid = 31794245 | s2cid = 208626110 }}</ref> All biochemical reactions are associated with known error rates and the more complicated the reaction the higher the error rate. Therefore, it is not surprising that the splicing reaction catalyzed by the spliceosome has a significant error rate even though there are spliceosome accessory factors that suppress the accidental cleavage of cryptic splice sites.<ref>{{ cite journal | vauthors = Sales-Lee J, Perry DS, Bowser BA, Diedrich JK, Rao B, Beusch I, Yates III JR, Roy SW, Madhani HD | date = 2021 | title = Coupling of spliceosome complexity to intron diversity | journal = Current Biology | volume = 31 | issue = 22 | pages = 4898–4910 e4894 | doi = 10.1016/j.cub.2021.09.004| pmid = 34555349 | pmc = 8967684 | bibcode = 2021CBio...31E4898S | s2cid = 237603074 }}</ref>

Under ideal circumstances, the splicing reaction is likely to be 99.999% accurate (error rate of 10<sup>−5</sup>) and the correct exons will be joined and the correct intron will be deleted.<ref>{{ cite journal | vauthors = Hsu SN, Hertel KJ | date = 2009 | title = Spliceosomes walk the line: splicing errors and their impact on cellular function | journal = RNA Biology | volume = 6 | issue = 5 | pages= 526–530 | doi = 10.4161/rna.6.5.9860| pmid = 19829058 | pmc = 3912188 | s2cid = 22592978 }}</ref> However, these ideal conditions require very close matches to the best splice site sequences and the absence of any competing cryptic splice site sequences within the introns and those conditions are rarely met in large eukaryotic genes that may cover more than 40 kilobase pairs. Recent studies have shown that the actual error rate can be considerably higher than 10<sup>−5</sup> and may be as high as 2% or 3% errors (error rate of 2 or 3 x 10<sup>−2</sup>) per gene.<ref>{{ cite journal | vauthors = Melamud E, Moult J | date = 2009 | title = Stochastic noise in splicing machinery | journal = Nucleic Acids Research | volume = gkp471 | issue = 14 | pages = 4873–4886 | doi = 10.1093/nar/gkp471| pmid = 19546110 | pmc = 2724286 }}</ref><ref name = Fox>{{ cite journal | vauthors = Fox-Walsh KL, Hertel KJ | date = 2009 |  title = Splice-site pairing is an intrinsically high fidelity process | journal = Proceedings of the National Academy of Sciences | volume = 106 | issue = 6 | pages = 1766–1771 | doi = 10.1073/pnas.0813128106| pmid = 19179398 | pmc = 2644112 | bibcode = 2009PNAS..106.1766F | doi-access = free }}</ref><ref>{{ cite journal | vauthors = Stepankiw N, Raghavan M, Fogarty EA, Grimson A, Pleiss JA | date = 2015 | title = Widespread alternative and aberrant splicing revealed by lariat sequencing | journal = Nucleic Acids Research | volume = 43 | issue = 17 | pages = 8488–8501 | doi = 10.1093/nar/gkv763 | pmid = 26261211 | pmc = 4787815 }}</ref> Additional studies suggest that the error rate is no less than 0.1% per intron.<ref name = Pickrell>{{ cite journal | vauthors = Pickrell JK, Pai AA, Gilad Y, Pritchard JK | date = 2010 | title = Noisy splicing drives mRNA isoform diversity in human cells | journal = PLOS Genet | volume = 6 | issue = 12 | pages = e1001236 | doi = 10.1371/journal.pgen.1001236| pmid = 21151575 | pmc = 3000347 | doi-access = free }}</ref><ref name = Skandalis>{{cite journal | doi=10.1016/j.mrfmmm.2016.01.002 | title=Estimation of the minimum mRNA splicing error rate in vertebrates | date=2016 | last1=Skandalis | first1=A. | journal=Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis | volume=784-785 | pages=34–38 | pmid=26811995 }}</ref> This relatively high level of splicing errors explains why most splice variants are rapidly degraded by nonsense-mediated decay.<ref>{{ cite journal | vauthors = Zhang Z, Xin D, Wang P, Zhou L, Hu L, Kong X, Hurst LD | date = 2009 | title = Noisy splicing, more than expression regulation, explains why some exons are subject to nonsense-mediated mRNA decay | journal = BMC Biology | volume = 7 | pages = 23 | doi = 10.1186/1741-7007-7-23| pmid = 19442261 | pmc = 2697156 | doi-access = free }}</ref><ref>{{ cite journal | vauthors = Bitton DA, Atkinson SR, Rallis C, Smith GC, Ellis DA, Chen YY, Malecki M, Codlin S, Lemay JF, Cotobal C | date = 2015 | title = Widespread exon skipping triggers degradation by nuclear RNA surveillance in fission yeast | journal = Genome Research | volume = 25 | issue = 6 | pages = 884–896 |  doi = 10.1101/gr.185371.114| pmid = 25883323 | pmc = 4448684 }}</ref>

The presence of sloppy binding sites within genes causes splicing errors and it may seem strange that these sites haven't been eliminated by natural selection. The argument for their persistence is similar to the argument for junk DNA.<ref name = Pickrell /><ref>{{ cite journal | vauthors = Saudemont B, Popa A, Parmley JL, Rocher V, Blugeon C, Necsulea A, Meyer E, Duret L | date = 2017 | title = The fitness cost of mis-splicing is the main determinant of alternative splicing patterns | journal = Genome Biology | volume = 18 | issue = 1 | pages = 208 | doi = 10.1186/s13059-017-1344-6| pmid = 29084568 | pmc = 5663052 | doi-access = free }}</ref>

<blockquote>Although mutations which create or disrupt binding sites may be slightly deleterious, the large number of possible such mutations makes it inevitable that some will reach fixation in a population. This is particularly relevant in species, such as humans, with relatively small long-term effective population sizes. It is plausible, then, that the human genome carries a substantial load of suboptimal sequences which cause the generation of aberrant transcript isoforms. In this study, we present direct evidence that this is indeed the case.<ref name = Pickrell /></blockquote>

While the catalytic reaction may be accurate enough for effective processing most of the time, the overall error rate may be partly limited by the fidelity of transcription because transcription errors will introduce mutations that create cryptic splice sites. In addition, the transcription error rate of 10<sup>−5</sup> – 10<sup>−6</sup> is high enough that one in every 25,000 transcribed exons will have an incorporation error in one of the splice sites leading to a skipped intron or a skipped exon. Almost all multi-exon genes will produce incorrectly spliced transcripts but the frequency of this background noise will depend on the size of the genes, the number of introns, and the quality of the splice site sequences.<ref name = Fox /><ref name = Skandalis />

In some cases, splice variants will be produced by mutations in the gene (DNA). These can be SNP polymorphisms that create a cryptic splice site or mutate a functional site. They can also be somatic cell mutations that affect splicing in a particular tissue or a cell line.<ref>{{ cite journal | vauthors = Scotti MM, Swanson MS | date = 2016 | title = RNA mis-splicing in disease | journal = Nature Reviews Genetics | volume = 17 | issue = 1 | pages = 19–32 |  doi = 10.1038/nrg.2015.3| pmid = 26593421 | pmc = 5993438 }}</ref><ref name = Shirley>{{ cite journal | vauthors = Shirley B, Mucaki E, Rogan P | date = 2019 | title= Pan-cancer repository of validated natural and cryptic mRNA splicing mutations | journal =  F1000Research | volume = 7 | pages = 1908 | doi = 10.12688/f1000research.17204.3| pmid = 31275557 | pmc = 6544075 | s2cid = 202702147 | doi-access = free }}</ref><ref>{{ cite journal | vauthors = Mucaki EJ, Shirley BC, Rogan PK | date = 2020 | title = Expression changes confirm genomic variants predicted to result in allele-specific, alternative mRNA splicing | journal = Frontiers in Genetics | volume = 11 | pages = 109 | doi = 10.3389/fgene.2020.00109| pmid = 32211018 | pmc = 7066660 | doi-access = free }}</ref> When the mutant allele is in a heterozygous state this will result in production of two abundant splice variants; one functional and one non-functional. In the homozygous state the mutant alleles may cause a genetic disease such as the hemophilia found in descendants of Queen Victoria where a mutation in one of the introns in a blood clotting factor gene creates a cryptic 3' splice site resulting in aberrant splicing.<ref>{{ cite journal | vauthors = Rogaev EI, Grigorenko AP, Faskhutdinova G, Kittler EL, Moliaka YK | date = 2009 | title = Genotype analysis identifies the cause of the "royal disease" | journal = Science | volume = 326 | issue = 5954 | pages = 817 | doi = 10.1126/science.1180660| pmid = 19815722 | bibcode = 2009Sci...326..817R | s2cid = 206522975 | doi-access = free }}</ref> A significant fraction of human deaths by disease may be caused by mutations that interfere with normal splicing; mostly by creating cryptic splice sites.<ref>{{ cite journal | vauthors = Lynch M | date = 2010 | title = Rate, molecular spectrum, and consequences of human mutation | journal = Proceedings of the National Academy of Sciences | volume = 107 | issue = 3 | pages =961–968 | doi = 10.1073/pnas.0912629107| pmid = 20080596 | pmc = 2824313 | bibcode = 2010PNAS..107..961L | doi-access = free }}</ref><ref name = Shirley />

Incorrectly spliced transcripts can easily be detected and their sequences entered into the online databases. They are usually described as "alternatively spliced" transcripts, which can be confusing because the term does not distinguish between real, biologically relevant, alternative splicing and processing noise due to splicing errors. One of the central issues in the field of alternative splicing is working out the differences between these two possibilities. Many scientists have argued that the null hypothesis should be splicing noise, putting the burden of proof on those who claim biologically relevant alternative splicing. According to those scientists, the claim of function must be accompanied by convincing evidence that multiple functional products are produced from the same gene.<ref>{{ cite journal | vauthors = Mudge JM, Harrow J | date = 2016 | title = The state of play in higher eukaryote gene annotation | journal = Nature Reviews Genetics | volume = 17 | issue = 12 | pages = 758–772 | doi = 10.1038/nrg.2016.119| pmid = 27773922 | pmc = 5876476 }}</ref><ref>{{ cite journal | vauthors = Bhuiyan SA, Ly S, Phan M, Huntington B, Hogan E, Liu CC, Liu J, Pavlidis P | date = 2018 | title = Systematic evaluation of isoform function in literature reports of alternative splicing | journal =  BMC Genomics | volume = 19 | issue = 1 | pages = 637 | doi = 10.1186/s12864-018-5013-2| pmid = 30153812 | pmc = 6114036 | doi-access = free }}</ref>