“…There are strategies to at least partially circumvent this problem, like using NNS instead of NNN codons (with N = A, C, G, T; S = C, G) taking advantage of redundancy of the third nucleotide positions in the majority of codons [ 10 ], or using spiked oligonucleotides [ 11 ], which are synthesized from solutions of the four nucleotide building blocks, each of those contaminated with a "spiking mix" consisting of equal aliquots of each of the four building blocks [ 9 , 12 ]. The required volume of the spiking mix to achieve a desired amount of nucleotide replacements at a defined position of the oligonucleotide can be calculated, such that library size and degree of randomization can be restricted [ 13 – 14 ]. Nevertheless, although those methods and sophisticated variations of them [ 14 – 17 ] have improved library design and synthesis, full control over randomization is not possible.…”