Abstract: ABSTRACT: BACKGROUND: Eukaryotic translation initiation factor 4G (eIF4G) is thought to influence the translational efficiencies of cellular mRNAs by its roles in forming an eIF4F-mRNA-PABP mRNP that is competent for attachment of the 43S preinitiation complex, and in scanning through structured 5' UTR sequences. We have tested this hypothesis by determining the effects of genetically depleting eIF4G from yeast cells on global translational efficiencies (TEs), using gene expression microarrays to measure the abundance of mRNA in polysomes relative to total mRNA for ~5900 genes. RESULTS: Although depletion of eIF4G is lethal and reduces protein synthesis by ~75%, it had small effects (less than a factor of 1.5) on the relative TE of most genes. Within these limits, however, depleting eIF4G narrowed the range of translational efficiencies genome-wide, with mRNAs of better than average TE being translated relatively worse, and mRNAs with lower than average TE being translated relatively better. Surprisingly, the fraction of mRNAs most dependent on eIF4G display an average 5' UTR length at or below the mean for all yeast genes. CONCLUSIONS: This finding suggests that eIF4G is more critical for ribosome attachment to mRNAs than for scanning long, structured 5' UTRs. Our results also indicate that eIF4G, and the closed-loop mRNP it assembles with the m7G cap- and poly(A)-binding factors (eIF4E and PABP), is not essential for translation of most (if not all) mRNAs but enhances the differentiation of translational efficiencies genome-wide.