How EasyPanel’s Automated Panel Design Algorithm Works?

Traditional Cytometry: How EasyPanel’s Automated Panel Design Algorithm Works?

EasyPanel’s proprietary algorithm incorporates the panel design rules below:

-Minimizing Spillover Between Fluorochromes (Based on either default theoretical Fluorochromes Normalized Emission Spectra or user-experimentally generated spillover spread matrices)

-Matching antigens to fluorochromes based on antigens’ expression levels and fluorochromes brightness. For example, low expressed antigens are matched to bright fluorochromes and vice-versa. 

-Matching antigens to fluorochromes based on antigens’ coexpression profile.

-Avoiding dyes susceptible to fixation in case the panel involves intra-cellular staining and/or fixation.

-Avoiding fluorochromes excluded by users and working around fluorochromes forced or fixed by users with specific antigens

-Mapping the optimized panel design to our expansive commercial product database featuring the catalogues of BD, Biolegend, ThermoFisher Scientific, Miltenyi Biotec, Beckman Coulter, Biotechne and many others…

Many panels are experimentally tested and validated by users: Feedback is constantly collected and incorporated into the algorithm to ensure continuous improvement.

• Minimizing Spillover Between Fluorochromes Based on Fluorochromes Normalized Emission Spectra

We developed a proprietary Python script that returns the combination of fluorochromes (of a given size, as requested by the user) with the lowest possible “total spillover score”, i.e., the lowest total sum of all spillover scores caused by “offender fluorochromes” into “offended fluorochromes”. 

A spillover score of an offender fluorochrome into an offended one is the ratio between the Area Under the Curve (AUC) of the offender over that of the offended in its corresponding optical filter region. 

For each panel of size R requested by users, the EasyPanel script:

1. creates different possible combinations (of size R) of fluorochromes including exactly 1 viability dye
2. calculates their respective Total Spillover Score
3. ranks the combinations by ascending order of Total Spillover Score
4. returns the combination with the lowest Total Spillover Score

Combinations containing spillover scores higher than a threshold of 55% between any 2 fluorochromes are rejected.

• Minimizing Spillover Spread Between Fluorochromes Using User-Imported Experimentally-Generated Spillover Spread Matrices

Alternatively to the 2 aforementioned scripts for traditional and spectral panel design, we developed a script for automated panel design based on experimentally-generated spillover spread matrices uploaded by users (using their own reagents on their own cytometer).

Spillover spread scores are scores varying from 0 to 1 throughout the spillover spread matrix (except for the matrix diagonal where the spillover spread scores between the same fluorochromes is equal to 1). 

The script minimizes the total spillover spread score (which is the sum of the spillover spread scores between the fluorochrome pairs of a given fluorochromes combination or panel). 

Fluorochrome combinations containing spillover spread scores higher than 0.7 between any 2 fluorochromes are rejected.

• Matching Fluorochromes to Antigens

We developed a proprietary script that matches fluorochromes returned in the optimized combination to antigens added by users as per the considerations below, ordered by higher to lower priority:

1. Antigens marked as co-expressed on the same cell population (cell population 1, 2 or 3) are matched to fluorochromes in “Fluorochromes Coexpressed group 1, 2 or 3”. In each of the 3 fluorochrome groups are 5 fluorochromes with minimal total spillover score (or minimal total similarity score or minimal total spillover spread score). 
2. Antigens not marked as co-expressed on the same cell population are matched to fluorochromes in the group “Remaining_FCs” (which are fluorochromes in none of the “Fluorochromes Coexpressed Group 1, 2 nor 3”) or to fluorochromes not yet matched (even if they are in the group “Fluorochromes Coexpressed Group 1”)
3. Antigens marked as “Low Expression Level” or “Unspecified Expression Level” are matched to the group of fluorochromes with the highest brightness scores (or stain index). Antigens marked as “High Expression Level” are matched to the group of fluorochromes with the lowest brightness/stain index.
4. Antigens marked as “Dump Channel” are matched to the Fluorochrome with the lowest brightness/stain index.
5. At this stage, further to all the above-mentioned matching criteria, specific matching of one antigen to one fluorochrome is not done yet, i.e., each antigen is matched to a group of fluorochromes, not a specific one.

Fluorochromes of the optimized fluorochromes combination are queried (vs each of the user-entered antigens and user-specified Species reactivity) in the commercial database to quantify and rank them in terms of “commercial availability”. The commercial availability score of each fluorochrome is the number of different antigens (in the corresponding matching group) for which the fluorochrome commercially exists.

Fluorochromes with lowest commercial availability score are first matched to “eligible” antigens.