Recombinant DNA Technologies

08/27/2026

R&D can be making real scientific progress and still be moving away from what the company needs.

Priorities change. Requirements change. New findings change what we know. Staying aligned means periodically checking whether the work still matches the destination.

The right priority today may be the wrong priority six months from now. Not because the project failed. Not because it was a bad decision. But because the assumptions that made it a priority changed. Six months ago, a company had 24 months of runway. There was enough time and money to pursue several...

08/25/2026

R&D can be making real scientific progress and still be moving away from what the company needs.

Priorities change. Requirements change. New findings change what we know. Staying aligned means periodically checking whether the work still matches the destination.

One of the hardest things for scientists to accept is that an interesting question isn't necessarily a useful question. New findings point in an unexpected direction. It's interesting, potentially important, and worth understanding. So you follow it. One experiment leads to another. Before long, you...

08/21/2026

Computational drug discovery has come a long way in 20 years. The models are dramatically more capable, but prediction still has to survive contact with the experiment.

The tools have changed. Biology still gets the final vote.

08/19/2026

Every screen has limitations. The key isn't always building a better screen - sometimes it's knowing which question each screen should answer.

A lesson from antimicrobial discovery that applies broadly across R&D.

08/17/2026

AI is changing the scale and speed of drug discovery, but faster screening doesn't eliminate one of the oldest challenges in R&D: Are you finding the right hits - and what might you be screening out?

A few thoughts inspired by a recent conversation with Akhila Kosaraju, MD, CEO and co-founder of Phare Bio.

If you're running a screen and concerned that you might be missing key hits, let's talk.

08/13/2026

FDAs recent actions are confusing, confounding and potentially dangerous. Read more here:

The FDA is proposing to replace so-called "self-GRAS" with mandatory filings. In and of itself, this makes some sense. The threshold for self-declaring your product as "Generally Recognized As Safe" is fairly low. But there is a tradeoff: cost. Building a solid GRAS filing to clear the "no questions...

08/12/2026

Today: Constraints help you make decisions.

In my last post I talked about how constraints define what "better" means in molecular diagnostics. The same principle applies when the assay looks completely different. - PCR: primers must coexist under the same reaction conditions. - Lateral flow immunoassay: antibodies, membranes, flow, signal, a...

08/10/2026

Today's article: Constraints tell you what matters.

Good science asks, "How well can we make this assay work?" Product development asks, "How well does this assay need to work within the constraints of the product we're actually trying to build?" At Curative, one of our key constraints was speed. We were developing a multiplex molecular diagnostic th...

08/03/2026

Great engineering isn't about optimizing a single variable. It's about understanding how every change affects the entire system. Here's Branden's perspective on balancing speed, sensitivity, and specificity in molecular diagnostics following last week's ADLM conference.

I talked to a lot of people developing molecular diagnostics at ADLM last week, and one theme stood out: speed is king. Almost everyone is trying to make molecular assays faster. But it’s a balancing act. I like to think of molecular diagnostics as having its own Iron Triangle: speed, sensitivity,...

07/29/2026

Day 1 (for me) at was great. On to day 2.

Day 1 of ADLM is in the books. It was great seeing many familiar faces and former colleagues like Ross Dehmoobed and Bruce Sargeant both formerly of Fusion Biotec. I worked with them when I was with Curative and I cannot recommend both of them strongly enough for your software (Ross) and engineering...