This technology uses novel synthetic 5’ unranslated regions (UTRs) to increase protein expression. 5’ UTRs are present in all genes and are required for ribosomes, the translation machinery that makes protein, to find, bind to, and begin translating RNA sequences. The ‘strength’ of a gene’s 5’ UTR dictates how much protein is made from that gene’s RNA, and 5’ UTRs from highly expressed genes have long been used in molecular biology applications. These inventors performed an unbiased screen of the 5’ UTRs on highly expressed genes across many cell types, trained a machine-learning algorithm with this data, then used this computational platform to develop efficient synthetic 5’ UTR sequences. The resulting synthetic 5’ UTRs drive higher expression than commonly used natural 5’ UTR sequences, therefore resulting in higher protein expression per cell.