How to align retention time

See Retention Time Alignment principles if needed.

Required condition to run the Retention Time Alignment algorithm:

• Selected User context must be grouped

If above condition is respected, then follow these steps:

1. Select the User context you want to align (You first need to group proteins for this context and save it)
2. To display the settings dialog:
   • Right-click on the selected context and select Algorithms→Retention Time→Retention Time Alignment
   • Or, from the main toolbar, select MS Identifications→Context→Algorithms→Retention Time→Retention Time Alignment
3. Change algo settings options (see below) if needed and click the Ok button

On top of the dialog, are displayed the following information:

• The reference User context
• The number of identifications to align onto the reference User context
• The directory where input/output files for NETPrediction utility are written (= the active hEIDI project directory)

Retention time panel

• Duration (min): analysis time in minutes
• Delay (min): delay in minutes

Outliers panel

• Don't take into account outliers: check this box if you want the algorithm to skip species with predicted retention time far from the others.
• Deviation threshold (%): if the above checkbox is selected, enter a thresh expressed in percentage. 100% represents the average absolute deviation (between RT & predicted RT). For example, a value of 200% means that species with predicted RT far about 200% from the average deviation are excluded from the calculation.

Grouping method panel

• Average: average the RT values to compute the representative RT value of a species group
• Median: use median to compute the representative RT value of a species group
• Cross-product: use cross-product to compute the representative RT values of a species group
• Combined: use median & cross-product to compute the representative RT value of a species group
• Best score: use average to compute the representative RT value of a species group

Performance panel

• Use memory tuning: if NOT selected, the process will be faster but memory consuming. If selected, the process will be a little bit slower but memory efficient.

When alignment process is finished, the following dialog message appears:

Each aligned identification has two new properties):

• slope: slope of the linear regression
• intercept: intercept of the linear regression
• reference: reference user context

You can access your identification properties either from identifications under your user context, or the under the appropriate search contexts. If you look at the identification properties just after running the Retention Time Alignment algorithm, it may not refresh immediatly.

To apply alignment coefficients on species, you need to Recompute Retention Time.

ATTENTION
During alignment process, retention time are predicted using an external utility and stored in peptide properties (peptides belonging to the reference context). These values are NOT the final recomputed retention times, they are only used by the retention time alignment algorithm.