I recently had a problem with protein purification in BL21, ran my sequence through a program linked to by a helpful Research Gate poster, and discovered that ~26% of my codons were reaching for tRNAs that represented <10% of the tRNAs for their respective amino acids

I recently had a problem with protein purification in BL21, ran my sequence through a program linked to by a helpful Research Gate poster, and discovered that ~26% of my codons were reaching for tRNAs that represented <10% of the tRNAs for their respective amino acids. I then ran my sequence through the optimizer linked below but have been checking the results manually as it is my first time and I'd like to understand better. Just a few of my codons have been switched for codons that actually represent fewer TRNAs for that amino acid than before. I am still at 0% below the above threshold overall but wanted to check in. The ones I'm seeing most frequently are: D: GAU (0.65) -> GAC (0.35) F: UUU (0.57) -> UUC (0.43) S: AGC (0.33) -> UCC (0.11) Is there something besides codon usage that is taken into account during optimization processes? Additionally. if a bottleneck is ultimately avoided is there any tangible benefit in using a more common codon?