Most eukaryotic genes are subject to alternative splicing (AS), which may contribute to the production of functional protein variants or to the regulation of gene expression, notably via 25 nonsense-mediated mRNA decay (NMD). However, a fraction of splice variants might correspond to spurious transcripts, and the question of the relative proportion of splicing errors vs. functional splice variants remains highly debated. We propose here a test to quantify the fraction of AS events corresponding to errors. This test is based on the fact that the fitness cost of splicing errors increases with the number of introns in a gene and with expression level. We 30 first analyzed the transcriptome of the intron-rich unicellular eukaryote Paramecium tetraurelia.We show that both in normal and in NMD-deficient cells, AS rates (intron retention, alternative splice site usage or cryptic intron splicing) strongly decrease with increasing expression level and with increasing number of introns. This relationship is observed both for AS events that are detectable by NMD or not, which invalidates the hypothesis of a possible link with the regulation 35 of gene expression. Our results indicate that in genes with a median expression level, 92%-98% of observed splice variants correspond to errors. Interestingly, we observed the same patterns in human transcriptomes. These results are consistent with the mutation-selection-drift theory, which predicts that genes under weaker selective pressure should accumulate more maladaptive substitutions, and therefore should be more prone to errors of gene expression.
40