I suggest running SDS PAGE with your recombinant scFv and simply look at the size of the protein. Alternatively you could blot it and incubate with an anti-heavy and anti-light chain ABs.

Depending on the bacterial strains you are using, there are permissive stop codons which some strains are translating into amino acids, so check the genotype of your bacteria. If you should be working with Tomlinson style libraries, there is e.g. a permissive stop codon between the ScFv and the phage envelope protein that is used for phage display. So one easily may switch between phage display and ScFv production by simply using different bacterial strains.

You also might want to check, where the diversity in you library is rooted. VH and VL are not completely randomized in my experience. So it might happen, that there is truncated protein that turns out to be a binder, but not a complete ScFv.

Secondly, never trust sequencing results. Esp. if there is something suspicious, have a look at the raw data and check, if the chromatogram is matching the text file. A nice tool for that is finchtv. https://www.ch.cam.ac.uk/computing/software/finchtv

Randy Chance Are those "stop codons" in-frame? a read-through?

Yuan-Yeu Yau There are stop codons within the VH sequence of all reading frames. Even the frame that showed the correct VL sequences had stop signals within the VH. So yes, I believe it was a read-through. I used DH5a e.coli for cloning and as Wolfgang Schechinger mentioned, it does have a supE44 mutation that may have caused it to suppress the stop signals. I will do as Agnieszka Szczepek suggests to see if there is any sign of VH present. However, even if it is present is it likely to be expressed in mammalian cells? I believe conventional cell lines may not have that 'stop-signal-suppressive' property?. On the other hand if the ELISA results were solely from the VL portion then does it seem reasonable to continue downstream experiments by amplifying that portion out and re-ligating into my vector? Thank you all for your inputs. Very much appreciated.

Randy, if you're planning to express your construct in mammalian cells, the stop codons will surely do their intended job. In addition, you won't get high expression yields (if any) due to codon usage bias. It's highly suggested to have the cDNA synthesized for mammalian use, most gene synthesis companies are offering free online tools for that. It's not that expensive, should be much less than 1000 USD for a ScFv now. This will save you a lot of nightmares and frustration. Don't forget to add the usual culprits (a promoter like CMV, a Kozak consensus element for a strong initiation of translation. Replacing the stops with what they actually have been translated to by the permissive bacterial strain) and replace the pelB bacterial signal sequence with sth that works for Eukaryotes, e.g. the IgG-kappa signal. Have a look at the pSecTag2 vector maps e.g. they will tell you what you need.

If your ScFv turns out to be truncated, maybe replace the stops in question with sth innocent (Gly, Ala, Ser Thr or similar) (Quick change protocol, homemade is fine, no pricy kit necessary, or (if restriction sites should permit) subclone the "good" parts into another ScFv plasmid, that doesn't have the stops, but gives you a full length canonical ScFv) and verify it works.

Good luck!