Cloudy V3.07: improved interface, new threshold finding options, and more

Rejoice! Increased user-friendliness and new analysis modes can be at your fingertips in a mere couple of clicks!

Whereas in previous versions a single channel of a single reaction had to used as input, these version will take the entire read.QX output object as input. A separate input variable 'well' allows for selection of which well(s) to analyse. Choice of the channel is done via the 'method' input variable. Under method, we now support analysis of either fluorescence channel, full duplex analysis (simultaneous analysis for both channels), and an evagreen-specific approach that corrects for channel cross-talk.

Another new feature of the V3 series is the option to choose different threshold setting approaches. Net to the standard 'cloudy' analysis, I have implemented the Generalized Extreme Value (GEV) approach of Trypsteen et al. on a single reaction basis (see the readme for some details on the implementation and link to their publication). Another threshold option available is 'mixed' where 'cloudy' and 'gev' are run sequential and the respective thresholds are averaged. As 'gev' tends to fail in reactions that are close to saturation, 'cloudy' is always used as back-up mechanism, ensuring valid output.

I now also offer a Shiny-app script that you can run locally to have a graphical user interface to analyze reactions with the 'cloudy' algorithm. In addition, a few plotting functions are available that should make inspecting reactions more straightforward.

Cloudy Version 2.08 & 3.03 release!

Rejoice! A new version of the cloudy algorithm for ddPCR analysis is available!

Version 2.08 has been released. This is intended as the final release of the V2 series of the algorithm. Few modifications have been made, mostly stability tweaks. This version is the evolutionary end-point of this form of the code. Extensive open-heart and bypass surgery was needed in order to accommodate new features. The result is the V3 series of the algorithm.

The main feature V3 adds is the ability to address cross-talk between the fluorescence channels in order to improve resolution. This is still under development but is currently working stably for Evagreen data. Expect future improvements and modifications.

The text of the read.me has been updated to provide some more background as well as instructions for use.

Enjoy.

Cloudy V2-05

Cloudy Version 2.05 has been released!

This version mainly contains further stability upgrades when dealing with significantly overlapping droplet populations (previous version threw an error and quit). Even in these difficult cases an output will be now be produced. In most cases this output should even make sense, but since the populations are severely overlapping, one should not expect the results to be accurate.

In addition, a manual threshold option is now available. When specified, the algorithm will use the user input as threshold rather than its own calculations. Other population parameters are still calculated. The threshold has also been added to the algorithm output.

V2.04 (as originally published) will remain available.

Version history:

• 01 added a step to determine the number of populations (eg: only one => abort, two => OK, three => warn and continue)
• 02 further fine tuned the 'only one population' procedure for the cases where there are very few pos or neg droplets
• 03 restructuring of the algorithm for stability and efficiency reasons
• 04 added control step for overlapping bounderies & use of minima as a general backup for non-sd based boundaries
• 05 added further robustness when dealing with overlapping populations, added manual threshold option, added threshold to output