Go Back

A Computational Analysis of the Effect of G Triplets on Splice Site Selection

Lior Pachter, John Dunagan, Daniel J. Kleitman and Bonnie Berger

Department of Mathematics and Laboratory for Computer Science, MIT

A recent paper by McCullough and Berget [1] discusses the effect of G triplets on splice site selection in short introns. Experimental evidence suggests that G triplets in introns play an important role in the definition of exon-intron borders as well as helping in overall splicing efficiency. In particular, it is shown that in a class of short introns, G triplets additively enhance recognition of the 5' splice site. Furthermore, the G triplets are necessary for correct splicing in the absence of a "good" pyrimidine tract at the end of the intron.

We discuss these results in light of computational work we have done examining relationships between G triplets, intron length, GC content and other parameters associated with introns. In particular, we establish that the number of G triplets downstream from the 5' splice site is strongly correlated with the appearance of a G in the first position after the GT consensus. This effect is beyond an overall correlation between that position and the GC content of the intron. Furthermore, we establish the somewhat counterintuitive result that regions rich in G triplets are correlated with ``good'' 5' splice sites (it has been suggested that some of the variability in vertebrate splice sites may be due to compensating effects such as G triplets). These findings also help explain the previously mysterious observation that a G in the first position after the GT consensus is strongly preferred over the consensus-matching A in the most strongly consensus-matching donor sites [2].

Finally, we analyze the relevance of these results to the problem of gene identification.

[1] A. J. McCullough and S. M. Berget, "G Triplets Located throughout a Class of Small Vertebrate Introns Enforce Intron Borders and Regulate Splice Site Selection", Journal of Molecular and Cellular Biology, Aug. 1997, p. 4562-4571.

[2] C. Burge and S. Karlin, "Prediction of complete gene structures in human genomic DNA", Journal of Molecular Biology, 268, p. 78-94.

Go Back