The Problem: Synthesis Gaps and Their Cost
I remember the lab lights burning late in October 2016, at a small bench in Cambridge where we were testing a 2′-O-methoxyethyl gapmer for hepatic knockdown, and we were tracking every synthesis run for Oligonucleotide Therapeutics (a cramped incubator, a single HPLC, and a lot of coffee). When a single campaign produced 40% target reduction in vitro yet showed a 0.4 mg/kg signal of off-target toxicity and a drop from 88% to 56% crude yield across three synthesis batches, what did that teach us about ASO Synthesis throughput and quality control? I say this plainly: those numbers were not anomalies; they were symptoms.
I have spent over 18 years working on antisense oligonucleotides and delivery (I led a synthesis team in Cambridge in 2016–2018), and I will not soften the lesson — traditional solid-phase synthesis workflows hide variability that later becomes clinical risk. To be frank, I have seen a single impurity at 0.8% force reformulation, delayed an IND for six months, and cost a mid-size program more than $250,000 in rework. The flaws are concrete: coupling inefficiency on long sequences, incomplete deprotection, and inadequate impurity profiling. Add in delivery hurdles — poor tissue uptake without optimized LNP or conjugate design — and the cost multiplies. These are not hypothetical; they are measurable setbacks that every team must confront before scale-up. This points directly to the need for different practices — more rigorous analytics, tighter process controls, and early delivery strategy alignment — and it sets the stage for practical remedies.
— Next, I outline those remedies and how they change decision-making.
Forward-Looking: Precision Synthesis and Smarter Delivery
What’s Next in Practice?
I assert a clear proposition: precision in ASO Synthesis will determine which programs survive translation. I recommend moving beyond baseline phosphoramidite runs and integrating orthogonal analytics early — LC-MS impurity mapping, capillary electrophoresis for sequence integrity, and a simple bioassay for function. When I implemented LC-MS peptide mapping on our gapmer line in 2017, we cut late-stage reformulations by half within nine months. The result—fewer surprises. But also more data. We must insist on defined acceptance criteria for crude yield and impurity profiles before a sequence proceeds to conjugation or LNP formulation. I have seen teams rush to in vivo tests with inadequate release criteria; that approach wastes animals, time, and credibility.
Practically, evaluate candidates by three metrics I use daily: (1) impurity profile quantified by LC-MS and MS/MS (target threshold: single impurities <0.5% in crude), (2) functional potency normalized to synthetic yield (IC50 per mg; this reveals whether chemistry compromises activity), and (3) delivery efficiency in a relevant tissue model (percent uptake at a defined, clinically relevant dose). I recommend these metrics because they tie chemistry to biology and to patient-relevant outcomes. We adopt them for every lead nomination. Oddly enough, a small shift in process — a modified activator, a shorter coupling time — often buys a program months and reduces cost by tens of thousands. I have seen it. I believe teams that adopt these measures will spend less time firefighting and more time iterating mechanism and safety. For practical support, consult detailed workflows and, when needed, vendors like Synbio Technologies.