Looking for ways to clean up your Metabolomic and other Omic data? Check out these tips.

Gretchen Lorenzi

Many researchers are aware that metabolomics is a finicky science with a multitude of factors that can affect data quality, but how many such factors are there? We’ve identified and optimized several per year over the past 15 years, and we keep finding new ones.

When the number reached 42—“the answer to life, the universe, and everything,” (Hitchhiker’s Guide to the Galaxy)—I hoped that would signal the end, but more factors kept coming (and still are). It seems there might be no end to the factors that dirty our metabolomic data.

Well I’m here to finally take a stand against dirty data in metabolomics (I’m a relentless optimist that way).

In that regard, this article launches a mission to help others save time and valuable resources while increasing the return on investment in metabolomic data. And these same tactics can also improve other omic data, including genomics, proteomics, and glycomics to name a few. I plan to create a series of articles to highlight some of the best routes toward better, cleaner data. My hope is that you’ll stumble onto a helpful gem or two along this path with me.

And to show you just how detailed you need to be, today I’ll kick off the 42 Factors Series with one of my favorite factors—the humidity in cell culture incubators.

Factor – Incubator Humidity

That’s right, if you grow cells, the humidity of the cell culture incubator is a variable that can affect metabolomic (and other omic) data quality. Pharmacologists have known it for years…low humidity contributes to evaporation of media from the edge wells of a multiwell plate, so the edge wells typically aren’t used. We were, nevertheless, confused when two members of our team acquired different results from identical metabolomic experiments conducted on two different shelves of our incubator. What could explain such a phenomenon? The two variables that came to mind were temperature and humidity.

That prompted us to use Track-It Data Loggers (Monarch Instrument; purchased from VWR Scientific) to monitor and record temperature and humidity throughout the interior chamber over the course of a few days. Unexpectedly, in our case, temperature was held sufficiently constant throughout the chamber, but a humidity gradient was detected:

Apparently the single water pan installed in our incubator caused the gradient; the lower shelves (nearest the water pan) were more humid than the top shelves. On the bright side, this factor was easy to optimize. Simply placing a second water pan on the top shelf eliminated the humidity gradient:

Take-home message: use a second water pan to eliminate humidity gradients in cell culture incubators and to improve accuracy and precision in metabolomics (and possibly other types of experiments).

It’s worth noting that our laboratory is located in Houston, Texas, USA, which is notoriously hot and humid outside with air conditioning running almost constantly. Some of our colleagues in Houston repeated our results in their laboratories, whereas others have failed to detect humidity gradients in their incubators. We haven’t conducted every exhaustive iteration of the experiment to test all potentially associated variables (e.g., water-jacketed vs. air-jacketed incubators), but we’ve conducted the experiment at different times of year (summer vs. winter) and obtained opposing results ourselves, suggesting that environmental variables such as air conditioning may be involved. Clearly, there is more to this story that we do not yet understand, but for the moment, we choose to err on the side of caution by including two water pans (one on the top shelf and one on the bottom shelf) in each incubator. For new-style incubators that supply humidity through an ultrasonically pulsed water supply, we find it helpful to keep the humidity at a high setting (95% recommended).

Am I blowing your mind yet? Well hang on tight. More to come.

Huge thanks to Cliff Stephan and his lab at Texas A&M for pointing this factor out to us, and to present and former members of my team, especially Wai Kin Chan, who helped troubleshoot this issue.

Do you have any experience with incubator artifacts? Any pressing metabolomics questions? Leave me a comment.

About the Author Phil Lorenzi

Share your thoughts

Your email address will not be published. Required fields are marked

{"email":"Email address invalid","url":"Website address invalid","required":"Required field missing"}

Free!

Book [Your Subject] Class!

Your first class is 100% free. Click the button below to get started!