Truncations, Side Products and Truth in Labeling for DNA-Encoded Libraries: A Cautionary Tale
Did you ever search a library catalog, dutifully note the location of a book that is allegedly not checked out, find its shelf space and discover that the volume is missing? And then notice that items nearby are misfiled?
The state of the art in DEL library creation is better than this. Skilled scientists in pharma and biotech are initiating projects with this technology widely now, aided in their explorations with varying degrees of support from DEL specialists like X-Chem. Large pharma clients may prefer to take the reins and go after their targets on their own, in private. But even these companies often take advantage of partnerships to avoid the time and expense of DEL library production.
What all clients should realize is that this technology is no longer experimental: You should expect precision and quality in the libraries and assistance from your partners.
Does Every DEL Project Require Feats of Valor?
I recently read an article that gave me pause: Selecting Approaches for Hit Identification and Increasing Options by Building the Efficient Discovery of Actionable Chemical Matter from DNA-Encoded Libraries.
The combined pharma/DEL authors do make many useful points. For instance, to perform an affinity-based screen, you need a high-quality, high-concentration target, and it’s important that the majority of it is appropriately folded and active. The authors introduce methods whereby you can show, for targets in an immobilized format, what fraction is capable of engaging with a relevant ligand. In fact, we take this kind of approach at X-Chem also.
But as the highly wrought title suggests, the hit identification exercise described in the article was less than straightforward. Yes, in the end, a crack scientific team achieved a positive outcome. However, in my view, an inferior library was the root cause of unnecessary complexity that negated some of the fundamental benefits of DEL and likely took extra months and needless heroics to overcome.
A Story of Mistaken Identity
When the team first synthesized the compound, in accordance with the scheme of the library and with the three co-enriched building blocks, that compound was inactive.
To understand what was going on, the team had to resort to on-DNA synthesis, a step that should not have been necessary. Utilizing building blocks A, B and C, compounds in this library should have reliably incorporated one of each. However, much downstream processing revealed that the most active compound did not conform to the intentions of the library design at all.
Not only was building block A absent in the active compound, but building block C reacted twice, a tendency that should have been detected during building block validation work.
Moment of Truth
The authors seem to conclude that this kind of predicament is just part of doing business in the world of DEL: “An often underappreciated challenge in the application of DEL hit ID technology is the complexity of chemistry that arises from the monomer diversity employed in the construction of the industrial DEL collections.” This complex chemistry then supposedly leads to “a poor conversion rate of NGS-identified hits to confirmed ligands off DNA, complicating the realization of value from the technology.”
I strongly disagree. Significant issues with incomplete yields, side product formation and other aberrations are the sign of poorly designed, poorly validated library schemes that will inevitably derail selection output analysis. Ambiguity in the final product simply makes a lot more work.
Some might argue that finding this off-scheme hit was serendipitous. But the cost in time and resources was too high. Why not just design the library so you get what you expect the first time? In this case, for one thing, the high reactivity of the B building block should have been recognized and addressed before the library was made. On-DNA follow-up should be unnecessary and really constitutes inefficient repetition of the original scheme.
The Moral: You Should Be Able to Rely on Your DNA Labels to Represent Your Original Intentions
The DEL platform is a powerful tool capable of simultaneously investigating a hugely diverse range of molecules that hold enormous potential. There are subtleties in its application that an experienced partner like X-Chem can help you sort through as needed. However, at this point, DEL projects should be relatively routine in that there should be a direct correspondence between the process history recorded in the DNA and the intended identity of the compound.
You should be able to rely on your library labels — that’s their whole point. At X-Chem, of the hundreds of compounds we’ve licensed over the last 10 years, 94% of them conformed perfectly to the design of the library. Is that record perfect? No. But it’s reliable enough to allow us and our clients to operate DEL as intended.
The X-Chem approach has always been to make discoveries with the original, affinity-based screen and then jump straight to activity assessment: a truly orthogonal assay. This approach is rooted in the fidelity of our libraries and underpins our 80% success rate in identifying functional hits, leading to 85+ licensed programs.