A separate tRNA for each of the 61 amino acids would incur significant costs for minimising translation errors and for maintaining each tRNA gene. Instead, the recognition accuracy of the third codon position is released to some extent, so that a single tRNA can couple to two or more codons by wobble base pairing.

A-to-I editing (deamination of A34 to inosine) occurs in all eukaryotes. The tRNAs with INN anticodons generated in this way can base-pair with C, U and A at the third codon position. All three and four-codon family boxes are usually decoded by an A-to-I edited tRNAINN and by both tRNAUNN and tRNACNN. Glycine codons are almost never decoded by tRNAACC, but by combinations of the other three tRNAs in that family box. tRNAGNN, tRNAUNN and tRNACNN are required to decode all two-codon sets. A tRNAANN in a two-codon set would be susceptible to A-to-I editing and subsequent mistranslating of codons of other two-codon sets.

The following genetic code matrix example shows whether all tRNAs required for decoding are present in an example genome assembly. tRNAs that are usually not present in four-codon boxes are highlighted in turquoise. The tRNA at the TGA stop codon decodes selenocysteine.