Request improved Precision on Retention Time Prediction Values

Request improved Precision on Retention Time Prediction Values Will Thompson  2023-02-14

Dear Skyline Team. Currently the max precision calculated for retention time prediction (reported in "Predicted Result Retention Time" field), is hundredths of a minute. For very fast runs with very narrow peaks, it seems like it would be useful to allow a more precise prediction. The current precision of 0.01 minutes is only 0.6 seconds, which for some separations is nearly the 4sigma peak width. Please see attached an example of a file where I think having a more precise result of the prediction algorithm would be useful for putting the prediction closer to the apex of the peak. Or please let me know if I've just not configured something correctly. Thanks for your help!



Nick Shulman responded:  2023-02-14
Thank you for pointing this out.

For most values in the Document Grid you can right-click on the column and choose "Format" and select a format that gives you more precision. Unfortunately, for the Predicted Retention Time and Predicted Result Retention Time, I see that the value is getting truncated much earlier in the process, so the Document Grid does not know what the full precision value is.

It would be straightforward for me to fix this in the Document Grid.

It looks like Skyline is also using the truncated value in intermediate calculations such as when calculating the "Retention Time Difference" score or when performing the linear regression between measured retention time and iRT score.
I will ask around and see whether we can fix it so that Skyline does not do any premature truncation of these retention time values.
-- Nick