Different 5′ untranslated regions (5′UTR) differ in their efficiency in attracting ribosomes to be positioned at the start codon. What Shine-Dalgarno (SD) sequences are optimal in prokaryotes? How are they decoded by the anti-SD (aSD) sequences? What is the most efficient Kozak consensus? How to characterize such a consensus for specific species, such as human or mouse, or for specific tissue where mRNAs are translated? How do signals, such as SD, coevolve with their decoders, such as aSD? How should bacteriophage and plasmid adapt their mRNA in response to their host translation machinery to achieve efficient translation initiation? How does translation initiation occur when the ribosome scanning is blocked by strong secondary structure at the 5′UTR? How do biopharmaceutical companies optimize their mRNAs to achieve efficient translation initiation? How do highly expressed genes optimize their translation initiation to achieve maximum protein production? How do rapidly replicating species, such as Escherichia coli, invest in their translation machinery differently from slowly replicating species, such as Mycobacterium tuberculosis? In particular, how does bioinformatics reduce all these complicated questions to numbers and provide numeric answers to these questions? Please read the chapter.