Protein translation is a fundamental, demanding process which requires several million ribosomes and consumes as much as 75% of cellular energy. Because of its importance, translation is regulated at many levels to maintain its high fidelity, with both substrates and synthesized products monitored by numerous quality control mechanisms. While post-translational quality control of proteins has been studied extensively, the mechanisms of co-translational quality control of nascent polypeptides, mRNA and the ribosome itself have only recently been appreciated. One of the major unanswered questions is how quality control mechanisms manage to specifically identify an aberrant event amid widely heterogenous normal physiologic states. This question forms the basis of this thesis, which is focused on understanding the molecular principles that determine accurate recognition of aberrantly slow ribosomes. To address this, we developed a novel flow cytometry-based assay to visualize terminal ribosome stalling at single cell resolution in mammalian cells. Using the assay, we firmly established that poly(A) messenger RNA (mRNA) is the most potent cause inducing terminal stalling. This system also led us to the identification and downstream characterization of a novel protein factor, the E3 ubiquitin ligase ZNF598, which we showed to be involved in triggering the quality control pathway during poly(A) translation. Subsequent in vitro ubiquitination experiments using purified ribosomes and ligase revealed molecular targets of ZNF598 - proteins eS10 and uS10 of the 40S ribosomal subunit. We further verified that ubiquitination of both targets is functionally important for poly(A)-mediated terminal stalling in cultured cells. Together, it led to the conclusion that ZNF598 recognizes excessively slow ribosomes and ubiquitinates them on the small subunit to initiate downstream quality control pathways responsible for the degradation of aberrant nascent proteins. To further understand how ZNF598 specifically recognizes and ubiquitinates aberrantly translating ribosomes, we reconstituted its recruitment to poly(A)-stalled translation complexes in an in vitro translation system. Unexpectedly, these experiments revealed that ZNF598 specifically associates with and ubiquitinates a sub-population of poly(A) stalled ribosomes consisting of closely-packed, collided di-ribosome species. This finding led to the general model of indirect detection of excessively slow ribosomes by ZNF598, whereby it recognizes ribosome collision events. We subsequently verified and generalised our proposed model using in vivo based experiments. In cultured cells, induction of stochastic ribosome collisions using sub-inhibitory doses of several unrelated translation elongation inhibitors led to the robust recruitment of ZNF598 to the sites of collisions, as manifested by ubiquitination of eS10. Our results explain the mechanism for sensing excessively slow translation at the molecular level. Moreover, the proposed model has also profound implications for general cellular physiology. Most importantly, the use of ribosome collisions to infer stalling means the degree of slowdown tolerated on an mRNA is tuned by the frequency of translation initiation; hence, the threshold for triggering quality control is necessarily substrate-specific.