A flow panel is all about choices. Our choices with the markers we include and the signals we attach to them via antibodies create a picture that helps us tell the story with our research. A well designed panel conveys that information clearly and helps us advance our science quickly and confidently. In this series of posts on panel design, I'll provide pointers from my experience with both conventional and spectral flow to help you avoid making the same mistakes and thus get better results quicker.
This first panel focuses on B and T cells markers relevant for interaction between the two cell types. It's good for telling stories about vaccine responses in lymphoid tissues.
This second panel contains some of the same markers, but incorporates the ability to differentiate various innate cell types and also look at senescence and polarisation in T cell responses. It's better equipped to tell us what's happening in the tissue during an active infection.
The first step in designing the panel is knowing what you want to stain for and why. You’ll get better results if your markers are informative for your experiments or patient cohort. If you include markers that aren’t useful or aren’t actually expressed on your cells, you’ll reduce your potential to get information.
Hopefully you’ve already reviewed the literature to know what markers should be present and which ones are considered important. As a second step, I recommend running some pilot studies or getting ahold of some existing high parameter data that covers your cells, conditions or markers. This is very helpful in understanding the shape of the data you’ll be working with. You need to know the expression levels of all the markers and which combinations are expressed together. Which ones change dramatically?
Why do we need to understand what the data will look like? We need this information in order to best pair our fluorophores with markers. At a basic level, we want to match low expression markers with bright fluorophores and highly expressed ones with dimmer dyes. Biology is messy and changeable, so low expression markers don't always stay that way. If we understand the potential range of expression, we can avoid blowing out the signal or creating unwanted noise. We also need to understand co-expression of markers and spreading between fluorophores to achieve the resolution we want. With a flow panel, we are not just measuring biology, we are reshaping the information, scaling it into an interpretable framework. This rescaling has important implications for the downstream data analysis, as we'll explore later.
One way of identifying potential range in your data is to look at the extremes of what you might be working with. In the example below, I was developing a panel for Tregs and ran a pilot panel on Treg-deficient mice with severe inflammation. There were some substantial changes that I wouldn’t have anticipated, so I redesigned the panel to accommodate those extremes of biology.
The backbone in this panel lacks all the activation markers, containing solely gating markers. This allows us to see where the negative ends and the staining begins for markers that don't always bifurcate neatly.
What happens if you don't do this? Well, if you plan to run tumour samples, but do all your optimisation on blood, you'll likely find that the panel becomes quite messy when applied to the tumour, where everything is upregulated. Similar effects can be seen if you compare lymphoid and non-lymphoid organs or even healthy and sick patients.
This knowledge of the biology is key to designing good panels. It allows you to get past basic aspects of panel design like pairing high expression markers with dim fluorophores and low expression with bright ones. There’s really no substitute for understanding what your data will look like.
I also recommend prioritizing your markers. A good flow panel involves compromise. Which markers do you need to resolve perfectly, and which are you okay with some loss on? If you need good resolution, you need to use a fluorophore that receives limited interference from others.
In the next post in this series, we'll look at the process of selecting fluorophores and understanding how best to pair those with markers in your panel.
Comments