Chemists can now make molecules faster than they can evaluate them. This imbalance defines early drug discovery. Automated synthesis platforms can generate thousands of crude mixtures in parallel, but analytical throughput still depends on gradients, equilibration, and compound-specific method tuning. Daniel Blair at St. Jude Children’s Research Hospital is well-acquainted with this issue.
“When you start to make even modest numbers of chemical reactions, studying them by LC–MS is prohibitive to progress”, observes Blair.
A single 384-well plate can tie up one LC–MS system for nearly an entire day. As reaction arrays expand, customization expands with them, and analytics begin to dictate which experiments chemists attempt.
Instead of forcing chromatography to keep up, Blair’s team reframed the problem. “We asked whether there are fundamental features inherent to every synthesis route that would let us analyze most reactions in the same way.”
Starting-Material Fragmentation as an Analytical Barcode
Most analytical workflows begin with a finished product, but Blair starts at the beginning.
Working in analytical science?
Register for a FREE Separation Science account to subscribe to the Separation Science Newsletter.
Under collision-induced dissociation, many starting materials break at weak, labile bonds. Those motifs often persist in downstream products. Rather than developing bespoke MS/MS transitions for every new compound, Blair’s group defines fragmentation behavior once, at the level of the starting material, and applies it across the analog series.
“Fragmentation is a fundamental feature of chemical matter, and we found that it is faithfully translated from starting materials to products,” he explains.
That fragmentation pattern becomes a barcode. The labile bond cleaves to yield a consistently neutral (uncharged) fragment. Any product that retains that motif registers through the same transition.
In Blair’s recent Nature study, this strategy appears as neutral-loss acoustic droplet ejection mass spectrometry (NL–ADE–MS). Acoustic droplet ejection delivers nanoliter aliquots into an open-port interface, and tandem MS (MS/MS) monitors a predefined neutral loss. One analytical definition anchors an entire reaction set.
Relocating Selectivity from Chromatography to Tandem MS
Once selectivity shifts into the mass spectrometer, the workflow changes.
“If you can introduce something into an MS/MS system at a pace of one per second, and you can derive an MS/MS profile from the starting materials, you don’t need the chromatography anymore”, asserts Blair.
With molecular specificity encoded in a predefined transition, runtime collapses to ionization and acquisition alone. ADE–MS delivers injections at roughly one per second. Throughput scales accordingly.
This architecture delivers immediate operational gains:
- Minimal carryover due to extreme dilution
- Low sample consumption
- Rapid replicate injections without gradient overhead
The approach is most effective when products retain the fragmentation motif; orthogonal methods are still needed when that motif is lost or when full structural assignment is required. LC–MS remains essential for confirmation and regulated work. However, reaction scouting no longer requires chromatographic separation.
Fit-for-Purpose Data: Ranking Over Absolute Quantitation
Early drug discovery asks a simple question: which condition is most effective? As Blair confirms, “You don’t need precise numbers—you just need to know which conditions perform best relative to the rest.”
By cutting acquisition time from minutes to seconds, experimental behavior shifts. Chemists can test more conditions, run more replicates, and narrow large design spaces faster.
Blair frames the value in economic terms. “The opportunity cost of doing these types of experiments quietly shapes which experiments scientists even attempt.” When analytics operate in seconds, exploration widens.
Tracking the Fate of the Starting Material
Fragmentation-first logic sharpens the analytical objective. “Your primary goal is to track the fate of that starting material. Where did it go and what did it do?”, emphasizes Blair. “Not all paths lead to product, and this approach lets you see where else that material goes”, he adds.
As the neutral-loss transition tags any species that retains the original motif, analysts can see incomplete conversion and competing pathways in the same channel. Full structural annotation can wait, but directional insight arrives immediately.
Standardized MS/MS as Infrastructure for Data-Driven Chemistry
High-throughput screening is meaningful only if the data remain comparable from plate to plate.
In conventional LC–MS workflows, gradients shift, retention drifts, and tuning decisions evolve over time. Individual datasets may be valid, but stitching them together across campaigns introduces variability that complicates comparison and modeling.
Fragmentation-first workflows reduce that drift by keeping the analytical method fixed while the chemistry changes. Each reaction is performed using the same MS/MS transition and acquisition parameters, producing rank-ordered outputs collected under uniform conditions.
“The limiting factor for AI-driven chemistry is no longer model design. What’s been missing is the ability to generate large amounts of data that are collected the same way, on the same instrument, and that you can actually trust”, explains Blair.
Under those conditions, reaction data accumulate as a coherent body of evidence rather than a patchwork of methods. Predictive efforts can focus on chemistry rather than on compensating for analytical inconsistency.
When Analytics Stop Being the Bottleneck
Fragmentation-first workflows do not eliminate LC–MS; they reposition it. Chromatography remains essential for structural confirmation, but early-stage reaction ranking no longer depends on gradients. At scale, the shift becomes structural.
“I would say the limitation now has become the pace at which we can make the chemical reaction mixtures,” concludes Blair.
When analytical architecture keeps pace with synthesis, reaction screening transforms from a necessary checkpoint into a strategic advantage. The faster chemists can see, the faster they can decide—and the faster discovery advances.
