November 15, 2016  |  Announcements, Release Notes, viSNE  |  By  |  0 Comments

New Release: viSNE Enhancements

Get Bigger, Better Insights in the Cloud with Our New Enhancements to viSNE.

Take advantage of the scalable compute available in the cloud with our latest enhancements to viSNE. Now Cytobank’s cloud-hosted platform can run up to 20x more events than other locally-run solutions. Fully optimize the resolution of your results by fine-tuning Iterations, Perplexity, and Theta. Each of these variables not only capture the dimensionality and detail in your results, they can also be used to optimize your workflows to achieve reliable results with the most efficient settings.

viSNE uses a step-wise procedure to optimize the placement of events in a two dimensional space that best reflects the similarity of events in the high dimensional space of the dataset. Each step of this process is an iteration of the algorithm. Each iteration progressively minimizes the difference between the high-dimensional similarity of the cellular events and the low-dimensional similarity of the events, which is displayed on the viSNE map. Iterations are critical for resolving the cellular populations on a viSNE map. The default number of iterations in Cytobank is 1000. The maximum iteration limit is now 7,500 for Premium accounts and 20,000 for Enterprise accounts.

Example of viSNE maps resulting from runs that did not have enough iterations to resolve the cellular populations. Each viSNE map shown here results from a separate dataset and viSNE run from either mass or flow cytometry data. Each viSNE map is colored by a major phenotyping marker such as CD3, CD4, or CD8. Each map shows a relative lack of resolution and spatial localization of cellular events that would make it difficult to use for subsequent gating or clustering to identify cellular populations.

Keep in mind that lowering the number of iterations will get results more quickly in situations where the algorithm can achieve adequate resolution of the cellular populations with fewer iterations. The flexibility of our new viSNE tools allow you to optimize the settings that work best for your dataset. Learn More

Like iterations, increasing perplexity can help you achieve resolution of the cellular populations in a viSNE map. Perplexity is used as part of calculating the high-dimensional similarity of two points before they are projected onto your viSNE map in low-dimensional space, and can be thought of as a rough guess for the number of close neighbors any given cellular event will have.  The default perplexity setting in Cytobank is 30, but the max value can be set as high as 100.

The effect of changing perplexity on different viSNE runs for the same dataset with all other settings being equal. Perplexity of 30 is usually a good default for cytometry data and identifies population centers with superior resolution than a perplexity of 10. Increasing the perplexity separates these population centers on the map more dramatically.

Since the original tSNE algorithm was limited in the number of events it could handle, viSNE in Cytobank uses an approximation to the original tSNE algorithm that makes it faster. The theta parameter can be increased or decreased to tune the balance of speed and accuracy in your viSNE run as compared to the original tSNE algorithm. A higher theta results in a faster run with coarser approximations of the points on the viSNE map (as compared to the map that would result from the original tSNE). In contrast, a lower theta results in a slower run with more accurate approximations of the points on the viSNE map. Theta = 0 corresponds to the exact version of the original tSNE algorithm, which can’t be run on datasets larger than a few thousand events. Generally, theta can be increased toward 1 and the viSNE results will still capture the cellular populations in the data. The default value or theta in Cytobank is 0.5. Learn More

The influence of different theta values on viSNE results. Theta = 0.01 can be considered the ‘gold standard,’ closest to what the original tSNE algorithm would produce. As theta increases from 0 to 1, the population centers remain fairly consistent and are good approximations of the original tSNE results. At theta = 1.5 (above the Cytobank limits for theta), the population centers on the viSNE map no longer look like the ‘gold standard’ theta = 0.01 population centers.

Get the Whole Picture with Our Newly-Expanded Color Palette
Easily visualize dozens of overlaid populations or other experiment variables with our new expanded palette of colors.
Use one visualization in the place where it would otherwise take multiple views by using color as an added dimension. Show context in a larger population, differentiate between populations, or reveal gating issues by using multiple dot overlays for up to 64 different colors.

Leverage Our Cloud Advantage
Regardless of the settings you choose, Cytobank’s hosted platform runs calculations in the background so you can continue to use your local computer for other tasks. In addition, you will always get a progress indicator to understand the pace of your run and the effects your chosen settings may have on ultimate run time. This not only helps you plan your workflows more efficiently, but also helps you understand the effects and tradeoffs between detail, accuracy and speed, so you can get the best results for your and your team’s needs.