Building a hypothesis-driven discovery method with multiple target monitoring (MTM)

In step 1, we used gas phase fractionation and information from a "pilot" experiment to create a Skyline document with a large number of potential targets. Here, we will discuss how to build a large scale targeted method that prioritizes potentially interesting biology to maximize the utility of an assay.

The priority file

Normally, PRM Conductor attempts to build an assay solely based on maximizing the number of unique, high quality proteins/ molecule groups it covers. However, in many cases there are specific protein or peptide targets that the user wants to include in their assay. Therefore, PRM Conductor has an optional input of a priority file (PF), which can be used to bias the assay towards desired targets.

What should go into a priority file?

Simply put, the PF should contain proteins or peptides (or precursors) that the user wants to include in the assay. There are many ways one could come up with this list. Some ways to generate a priority list include: 
  1. Features that look interesting in a pilot experiment on small number of samples 
  2. Targets that come up from a literature search/ past experience with a system
  3. Specific pathways that are known to be interesting
  4. Proteins that are predicted to be differential/ involved in system
  5. Known targets of drug
In this walk through, we are analyzing a 3-proteome mix. In this system, we know that yeast and E. coli peptides should be changing. Therefore, we will add all yeast/ E. coli peptides to the priority file

How PRM Conductor uses the priority file

In short, PRM Conductor prioritizes adding targets from PF to the assay. Practically, depending on the size of the list and the assay capacity, this can look different. In the case shown here, there are 5.7k peptides on the priority file. This is more than we can feasibly target in this assay. Therefore, PRM Conductor will try to add as many as it can from the prioritized list to assay. In the case of normal PRM, this is straightforward and will result in the assay consisting solely of prioritized features. However, with multiple target monitoring (MTM), there will be a mix of prioritized and non-prioritized species based on how the instrument can maximize coverage. 
 
In the case where all items in the PF can be monitored, PRM Conductor will add them and fill in the rest of the assay to maximize coverage. It is important to note here that if protein names are used, PRM Conductor will add all the peptides from those proteins, regardless of whether they passed the Refine Targets filters and ignoring the max peptides per protein selection. In the case where multiple assays are generated (balance load is unchecked), items from PF will be added to every assay. This makes it a good way to specify QC peptides.

Generating a priority file 

The PF is simply a plain text document where each line is either a peptide sequence or protein name and saved with the extension .prot. The entries must exactly match the protein/ molecule group name or peptide modified sequence exactly as given in Skyline. It is therefore suggested that these names be taken directly from the Skyline document grid. The document grid can be exported as .csv, filtered by any criteria or lookup function, and then the relevant column can be copied into a text editor and saved with the .prot extension.
 

Building targeted assay with priority file

We will now pickup where we left off in step 1. To build the assay, it is preferable to have only one gas phase fractionated library run. Therefore, delete the other two library runs from document, and open PRM Conductor with just the single library run loaded. PRM Conductor should remember the refinement setting we already used, so there is no need to touch the "Refine Targets" box. One could even check "Keep All Precs." to ensure that the document isn't further filtered. We will focus attention on the Define Method box for this step.
 
 
Here is a list of parameters under define method that can be selected, a short description, and suggested starting points
Parameter Description Suggested starting point 
Analyzer Impacts acquisition constraints depending on analyzer abilities Always select ion trap for Stellar MS
Scan Rate (kDa/s) Rate at which ions are ejected from trap, faster rate means lower resolution  Typically use 125 kDa/s. For higher resolution, slower scan rate can be used, but this may slow down instrument
Min Dwell (ms) Minimum accumulation time for a precursor in the assay  5 ms for high load applications. Can be increased up to 1000 ms when ultimate sensitivity is needed
Acquisition type What kind of experiment is done. MS2 is regular PRM. MS3 can be used in cases where selectivity limits performance. MTM increases MS2 assay capacity by combining windows. dDIA, dynamic DIA, creates time shifted DIA method, as described by Heil et al.  MS2 or MTM
Scan Range Mode What m/z range the MSn spectrum will cover. Larger range means slower acquisition rate.  Optimize automatically adjusts the scan range to cover the minimal range that includes transitions to be monitored 
LC Peak Width (s) Average width of LC peak. This is auto-populated based on observed peak widths in the document Default value (set based on peak width in data)
Min. Pts. per Peak Minimum data points per average LC peak  8-12
Cycle time (s) Peak width / desired points per peak  Cannot be adjusted. Calculated based on previous two values
Acq. Width Type Whether to use the same acquisition width for all targets (global) or to set acquisition width based on observed peak width Either works, if there is a lot of variance in peak with then per precursor should be used 
Acq. Window Either length of time to acquire all peptides for (if global width type is used) or the factor that peak width is multiplied by to determine per precursor acquisition width At least 3x average peak width (or factor of 3). Can be widened if chromatography is especially unstable 
Acq. Window Optimize Whether or not the assay should extend acquisition bounds beyond minimum width as defined above when there is time to do so. This is especially useful as RT can shift at beginning of run and Adaptive RT may not always account for these early shifts On
Max Precs./ Group Max molecules or peptides per group/ protein 5-10
Priority file Described above To use, double click to open a window to navigate to the .prot file
 
 
To start, we should select acquisition type. MS2, MS3, and MTM are all compatible with this workflow. To maximize total assay coverage, we will be using multiple target monitoring, or MTM, for this tutorial. Select from drop down menu and wait for PRM Conductor to perform preliminary processing and save metadata file related to MTM analysis.  Subsequent launches of PRM Conductor and selecting of MTM for this document will not take as long as the first time you select MTM.
Once MTM is selected, several new parameters will appear as well as a second window showing information about assay 
MTM min width is minimum isolation window size. If Opt. is checked, then window widths will be determined such that precursors are at least 1/2 the min width distance from the edge of the isolation window. If it isn't checked, then the same windows from DIA discovery run will be used. Max. Targets/ Acq. refers to the maximum number of targets in a single isolation window. Limiting this to a smaller number reduces concerns about dynamic range being hurt by co-isolating too many precursors. The plot that opens in a new window displays isolation widths used in the assay.
 
For the rest of the assay, we will use the following parameters. This includes selecting the priority file that was generated for all yeast and E. coli peptides in list, EcoliYeastPeps.prot: 
The text to the right of priority file indicates that we are acquiring 2779 of the prioritized features. The final assay should look something like this, with 3.6k precursors acquired in 1.7k different acquisition. Note that there may be some variance depending on a few different factors. This variance is largely because there are so many ways that this assay could be built.

Exporting MTM method

Before exporting the method, we need to create a template. Th easiest way to do this is go to GPF template and replace the DIA experiment with MSn (but keep the Adaptive RT DIA experiment). Most things will autopopulate when the method is exported, but make sure to use the same CE as the GPF method, and turn on Adaptive RT for Dynamic Scheduling (no need to include reference file): 
The template method can be found in data folder, 28min_PRMTemplate.meth.
 
Now, double-click the Method Template box to choose the template method and export the method with the Export Files button. You can either check the box to make a new Skyline document or click "Send to Skyline" and then later resave the Skyline document with a new file name.
Once the process is finished, you should check the method file to make sure it is as expected. If you look at the isolation list, you should see that different window widths are used and a RT Bin file has been added: 
 
Now you should be ready to acquire data!
 

Importing MTM data into Skyline

Next, we need to import data into Skyline. I recommend starting with Skyline document, 3Proteome_28min_MTMAssay_NoResults.sky, to make sure that precursors match exactly with tutorial data. To start, make sure GPF data is imported (if you exported a new Skyline file, you will need to re-import after changing the isolation scheme to 1 Th DIA.
 
To import MTM data, we need to change the acquisition scheme. We cannot use PRM as the acquisition scheme because precursor masses won't line up with center of isolation window in many cases. Therefore, we make a custom acquisition scheme. Go to Settings > Transition Settings and under Full Scan, select DIA as acquisition method. Under isolation scheme, select Add: 
Create a new isolation scheme called MTM as follows:
Click Ok, then select isolation scheme from drop down menu. Before leaving this page, increase retention time filtering window to 2 minutes in case there has been RT drift: 
Click over to instrument tab in transition settings. Ensure you still have the 40 Th scan filter applied and check the box for triggered chromatogram acquisition. Checking triggered chromatogram acquisition ensures that if there is a gap in acquisition (i.e., a precursor falls into 2 different acquisition windows that acquired at different times with a gap in acquisition) Skyline won't connect the lines between the two acquisition blocks.
Click Ok. Save document. You should be ready to import data by going to File > Import Results and importing results.

Expert review of peak picking

An in depth tutorial on Expert Review is available now, but here is a quick overview. Expert Review significantly improves peak picking and is highly recommended. Here, we will use that same GPF library run as our reference. After MTM data has finished importing into Skyline, save the document (or open 3Proteome_28min_MTMAssay_Results.sky). Then open Expert Review tool. Click to select current document, select the chromatogram library run as reference, and click start:
Following peak picking, click to send to Skyline:
Congratulations! You have now created an MTM method and imported data into Skyline! Further processing can be done, check out Skyline tutorials such as Processing Grouped Study Data.