Trends in Genetics
Volume 20, Issue 2, February 2004, Pages 68-71
Journal home page for Trends in Genetics

Genome Analysis
How prevalent is functional alternative splicing in the human genome?

https://doi.org/10.1016/j.tig.2003.12.004Get rights and content

Abstract

Comparative analyses of ESTs and cDNAs with genomic DNA predict a high frequency of alternative splicing in human genes. However, there is an ongoing debate as to how many of these predicted splice variants are functional and how many are the result of aberrant splicing (or ‘noise’). To address this question, we compared alternatively spliced cassette exons that are conserved between human and mouse with EST-predicted cassette exons that are not conserved in the mouse genome. Presumably, conserved exon-skipping events represent functional alternative splicing. We show that conserved (functional) cassette exons possess unique characteristics in size, repeat content and in their influence on the protein. By contrast, most non-conserved cassette exons do not share these characteristics. We conclude that a significant portion of cassette exons evident in EST databases is not functional, and might result from aberrant rather than regulated splicing.

Section snippets

Aberrant alternative splicing

Somatic mutations within splice sites or introns could result in aberrant splicing, leading to non-functional mRNAs; ESTs derived from these mRNAs would be indistinguishable from normal splice variants. Because somatic mutations are prevalent in cancer related tissues, and >50% of the ESTs in dbEST come from cancer, cell-lines or tumor tissues [6], such spurious variants can be ubiquitous in dbEST.

Splicosomal mistakes have also been proposed as a mechanism that can result in non-functional

Detecting cassette exons conserved between human and mouse

We have recently collected a dataset of 980 EST-predicted human alternatively spliced cassette exons [7]. From these 980 exons, 243 (25%) were also found to be alternatively spliced in mouse [‘conserved alternatively spliced exons’ (CAS exons)]. The remaining 737 (75%) are ‘non-conserved alternatively spliced exons’ (non-CAS exons) (Box 1). Low levels of alternative splicing conservation between human and mouse were also observed in other studies 8, 9. The method that was used to locate

Comparing conserved with non-conserved cassette exons

Presumably, orthologous exons that are alternatively spliced both in human and mouse have functional importance. We can therefore regard the group of CAS exons as a representative group of functional, alternatively spliced exons.

Non-CAS exons, on the other hand, can also be functional, representing exons created after the divergence of the human and the mouse lineages. However, if these exons, as a group, were indeed functional, we might expect them to have the same general characteristics as

Frequencies of ESTs, repeats and size as measures of function

It was recently suggested that a higher number of ESTs/mRNAs supporting a splice variant correlates with its functionality [11]. Our results agree with this observation: although CAS exons were on average supported by nine sequences (median 3), the average EST and mRNA support for non-CAS exons was 2.2 sequences (median 1). However, sequence support by itself is not sufficient for detecting functional alternative splicing: in our study, 30% of the CAS exons were supported by a single human

Which exons are functional?

We detected a conserved mouse exon for 25% of the human cassette exons in our set. These CAS exons most probably have functional importance. In principle, some of the 75% human non-CAS exons could also be functional; however, these might be expected to have similar characteristics to those of the CAS exons. This is not the case. We have shown that the group of non-CAS exons significantly differs from the group of CAS exons in many important parameters (discussed previously) (these differences

Comparison with other results

We examined the results of Modrek and Lee [9], who compared exon-skipping events between human and mouse using ASAP, an EST-based alternative splicing database. 127 ASAP exons were identified in which both variants (exon inclusion and exon skipping) were observed in human and in mouse (equivalent to our CAS exons); in 78 (61%) the length of the exon was a multiple of three (i.e. did not cause a frame shift). By contrast, of the 427 ASAP human exons that were predicted to be skipped in human but

Acknowledgements

We thank Dan Graur for inspiration; Han Xie, Amit Novik, Dvir Dahary and Ami Haviv for insightful discussions; Zurit Levine for technical help; and Eli Eisenberg for critical reading of the manuscript.

References (21)

There are more references available in the full text version of this article.

Cited by (252)

  • Bioinformatics approaches for studying alternative splicing

    2018, Encyclopedia of Bioinformatics and Computational Biology: ABC of Bioinformatics
  • Global regulation of alternative RNA splicing by the SR-rich protein RBM39

    2016, Biochimica et Biophysica Acta - Gene Regulatory Mechanisms
  • Two alternative splicing variants of a sugar gustatory receptor modulate fecundity through different signalling pathways in the brown planthopper, Nilaparvata lugens

    2019, Journal of Insect Physiology
    Citation Excerpt :

    In non-model insects, such as N. lugens, fewer AS variants have been identified (Kang et al., 2018). In addition, many AS variants may have no function; for example, approximately 75% of the detected AS in humans is likely not functional (Sorek et al., 2004). To our knowledge, NlGr10 is the only sugar gustatory receptor whose AS variants have been identified in N. lugens (Kang et al., 2018).

View all citing articles on Scopus

Supplementary data associated with this article can be found at doi: 10.1016/j.tig.2003.12.004

View full text