“We should’ve just done a DOE" ??
Greetings!?When it comes to Design of Experiments (DOE), here at Synthace we try and practice what we preach. But sometimes without realizing it, old habits can creep back in.?
They certainly did when we were developing our miniaturized purification product. We were running anion exchange chromatography experiments to purify individual proteins from a protein standard mixture using gradient elution.
It was our first time trying this kind of purification method, and somehow we found ourselves defaulting to running a bunch of OFAT experiments—the classical kind where you vary one factor at a time.?
We used some set points identified from the literature, but the results we got were nothing to write home about.
Then, we started exploring some factors. We tweaked the pH of the buffers, then the amount of salt in the buffers, then the size of the gradient. We kept going and before we knew it, we’d run around 10 OFAT experiments.
We’d achieved the nice separation of proteins we were after, but we hadn’t learned much about how our factors interacted with each other.
?“We should’ve run a DOE”—my colleague grumbled with an exasperated sigh, as she was writing up all the experiments into a single report.?
Of course, OFAT has a place in every scientist’s arsenal of methods. Very occasionally, you’ll already know that there's no chance of factor interactions.
More likely, there's just one thing you need to learn the right limits for, before you can plan the full experiment (things like temperature, enzyme concentration and cell number can often fall into this category).?
But the problem arises when you keep using OFAT as your main experimental strategy.
Our experience reminded us the hard way that whether you’re using OFAT or DOE, each experiment must have a clear purpose. It’s all too easy to fall into the trap of OFAT, even when it doesn’t actually suit your needs—so avoid being its next victim!??
'Til the next one,
?? DOEs have to be a compromise
If you're struggling to imagine how you'd apply DOE to the delicate business of biology, this one's for you. Our Global Manager of Field Application Scientists Ross Kent talks about balancing the wet, squishy nature of R&D lab work with statistical best practice.
?? "Why's my robot doing a little dance?"?
Disappointingly, Customer Success Scientist Nathan Hardingham's Tecan Evo didn't spontaneously decide to have a boogie. Though it was comforting to hear that the movement was intentional—and necessary for reducing sample waste and cleaning in subsequent steps.
?? The many AI models of drug discovery ??
Our Chief Scientific Officer Markus Gershater reminds us that accelerating drug discovery with AI doesn't mean waiting until we've brought all our data together. There are plenty of business-critical functions that need their own smaller, highly focused datasets and AI models. The best bit? We can start building those datasets with automated, high-dimensional experiments right now.?
领英推荐
?? Content we're loving?
Why we loved it:
It looks like we're one step closer to better understanding those who are suffering from long COVID.
Researchers who performed phenotypic profiling on patients hospitalized with the condition discovered that in different subtypes of long COVID, unique inflammatory pathways are being activated.
This means that there are different biomarkers that can be associated with distinct types of long COVID, which could be used as future therapeutic targets.
Why we loved it:
Such an?exciting breakthrough for Alzheimer’s research!
Researchers have developed an innovative neuronal model to help study the spreading of tau—a protein that accumulates in the brains of people who suffer from age-related neurodegenerative diseases like Alzheimer’s.
So far, there have been limited human models available to study these diseases. To remedy this, they edited human pluripotent stem cells using CRISPR technology, allowing them to express tau protein associated with ageing brains.
This hasn't just allowed them to simulate tau build-up, a process that would normally take decades in the human brain: They've also identified a number of novel gene targets which may be involved in tau spread.??
Why we loved it:
It was great to see UCL partner with some of our super talented?scientists at Synthace.
Together, they developed a systematic framework to increase the efficiency of aqueous 2-phase extraction (ATPE) process development, an essential purification step in bioprocess development. Using mathematical modelling and automation, they explored the relationships between different liquid transfer parameters and process error.
This helped them improve the accuracy and precision of liquid handling in ATPE, based on the mechanical properties of liquids, such as viscosity.
About Synthace
Get faster, smarter insights from your R&D experiments. Designed by and for biologists, Synthace lets you design powerful experiments, run them in your lab, then automatically build structured data. No code necessary. Learn more.
Get the latest from our lab team
Scientists Luci and Fatima keep you posted on DOE, lab automation, plus all things current and future-facing in life sciences R&D. No frills. Only the important bits.