The sequence of entire genomes from higher eukaryotes revealed that an average gene is very large due to the interruption of the coding sequence with large noncoding introns. Introns are co-transcriptionally removed by splicing with great accuracy and fidelity, although contrary to our expectations, currently known signals required for pre-messenger RNA (mRNA) processing are very degenerate and redundant. Furthermore, the vast majority of genes are alternatively processed. A large number of proteins are therefore involved in generating specificity in pre-mRNA processing that requires a dedicated mechanisms to operate at genomic dimensions. In this review I will summarize recent progress in understanding how established principles of pre-mRNA processing extend to genomic dimensions and discuss emerging concepts in coupling of pre-mRNA processing with other nuclear events and nuclear organization.