|
Dear Skyline Users, During our second Skyline Tutorial Webinar we received many great questions from our the participants -- here are the answers! Q: Shouldn't DIA be insensitive to RT? Ans: The acquisition side of DIA is independent of any expected peptide retention times. Unlike SRM and PRM, no matter how many peptides you may want to interrogate during data analysis, no scheduling is required in DIA. For data analysis, however, it is very important to know the retention times at which you expect to see peptides elute. Having this information greatly aids in correct peptide chromatogram peak assignment, and therefore accurate quantification. Many successful DIA chromatography-based quantitative analysis methods today rely on prior knowledge of peptide retention time. Though that knowledge can be gained after the DIA data is acquired, and applied to new analysis of existing data sets. Q: What about relative product ion intensity differences between DDA and DIA, which was described in the recent Aebersold article? Can you accommodate some difference in their ratios? Ans: In general, as described in the session, we expect relative product ion abundance from run-to-run on the same instrument to be very consistent. Different instruments of the same type should also be very consistent. Even MS/MS spectra from DDA data compared with peak areas from DIA data on the same instrument or multiple instruments of the same type should be very consistent. Though, here we expect more variance in the relative ion abundance measures taken from DDA MS/MS spectra, since they rely on fewer ions, potentially far fewer, if the spectrum was taken early or late in the peptide elution profile. To combat this problem, the Aebersold lab suggests using consensus spectra for calculating relative ion abundance from DDA data where you have multiple spectra for the same peptide. We expect, however, that the most reliable library building approach for relative product ion abundances in DIA will come from using extracted ion chromatogram peak areas from DIA data itself on the same type of instrument as will be targeted in your experiment. Next best, is probably the consensus spectrum approach proposed by the Aebersold lab. Reliability will deteriorate from there with difference in instrument types, etc. Q: Is it possible to use DIA data for identification -- as we do for DDA data -- by using a traditional protein search engine? Ans: As mentioned in the webinar, we know of 4 groups (MacCoss, Smith, Nesvizhskii and Bandeira labs) actively working on peptide detection software for DIA data that rely on no prior knowledge of fragment ion relative abundance or retention time. These project are still quite young compared with DDA peptide ID software, but we believe they show promise, and the Skyline team is actively working with all of them to ensure that the workflows that will involve these new tools integrate well with quantitative analysis of the DIA data. Also, Waters has long supported peptide searches of their MSe data, which is a type of DIA, and they even now support exporting pseudo-MS/MS spectra that can be searched with traditional peptide spectrum matching tools. My understanding is that at least the tool being developed by the Nesvizhskii lab for modern DIA on non-Waters instruments has this capability as well. Q: Is it possible to do DIA with a LTQ-Orbitrap instrument? Ans: Yes, it is possible, but this is a slower scanning instrument than the Q Exactive or modern TOF instruments, which means you will need to sacrifice either cycle-time or measured range to use the 20 m/z windows demonstrated during the tutorial. Q: Can DIA experiment be done on Orbitrap XL Discovery where the max. resolution is 30K? Ans: Yes. It is not necessary to have greater than 10K resolving power to perform DIA. The wider the resolution, however, the less selective your method will become. Also, as described above a slower scanning instrument will require sacrifices in either measured range or cycle time to use the same window size, and as resolution decreases, it may be necessary to decrease the window size to maintain acceptable selectivity. All of these factors are in competition for the quality of your digital record of all ions. As scan speed and resolution increase your digital record will improve in data quality. Q: Can DIA be done on ABSciex API - 4000? Ans: No. This is a triple quadrupole mass spectrometer. Although it has the ability to measure full-scan mass spectra it is not likely that they are fast enough or high enough resolution to support DIA. This instrument is best suited for SRM, where Skyline should still be a great help. Q: Is the isolation window export supported for AB Sciex 6600 TripleToF? Ans: Not in version 2.6, but AB Sciex has agreed to fund development effort in Skyline starting in January 2015, and our stated list of goals for that effort include isolation window export for AB Sciex instruments. In the current implementation, however, you would design your isolation scheme with AB Sciex software, and then port it into Skyline using the isolation scheme editing form presented in the tutorial. This form is extremely flexible, and we believe it supports all isolation schemes currently in use on AB Sciex instruments. Q: Are you planning to extend Skyline to longer peptides? Middle-down proteomics requires 50-60 amino acid long sequences, and I see this is above the limit accepted. Ans: I able to insert a peptide of length 175 with charge states 15-19 (m/z 1305.86 - 1031.15) using the Edit > Insert > Peptides form. It is a bit unwieldy in the Targets view to have such a long peptide, and I needed to add the charge states 15-19 to the list of precursor charge states in the Transition Settings - Filter tab, but it does appear to me to be possible. Others have asked for better support for full top-down chromatogram extraction with charge states up to 50. It is likely that we will make this easier in the future, but it is certainly possible up to charge state 20, right now, which is the only limit currently in place that should limit use of longer peptides. When you try to paste peptide lists with longer than 60 or so amino acids directly into Skyline, without using the Edit > Insert > Peptides interface, then Skyline may interpret these to be protein sequences, and ignore the line breaks to treat everything as one FASTA sequence. But, you can easily get around this with the Insert Peptides form. Q: When making a DIA method in Skyline, how do you set up overlapping between windows? (Say I want to scan 399-420, 419-440, etc.?) Ans: The important thing to realize is that you also want to include a margin with this type of isolation scheme. That is, the instrument may be collecting 399-420 and 419-440, but you only want Skyline to extract chromatograms from 399.5-419.5 and 419.5-439.5. I hope we have made this easier to understand in version 2.6, where we have made it possible to view either "Measured" windows or "Extraction" windows when using a margin, where the former is what you specified, and the latter is what I have just explained. You would start by specifying in the isolation scheme editing form that you want to include a margin. In your case, as in the SWATH method described in the Gillet paper, you would set that margin to 0.5 m/z to exclude from chromatogram extraction a range where the Gillet team measured signal drop-off on the AB Sciex 5600. Then you could choose which type of window ranges you prefer to enter from a dropdown list that appears when you specify the use of a margin, either "Measurement" or "Extraction", and you can toggle back and forth to see the difference, as I have explained above. Finally, you can always double-check your isolation scheme with the Graph button, which will show you the overlap and the margins you have specified, and warn you if you have accidentally specified a small overlap in the extraction ranges. Q: is it possible to set first mass to 0 in the MS2 parameter settings (Qex)? Ans: We tried setting the first mass to 0 and found that the lowest value the method editor would take was 50 m/z. Q: What is the point of using contaminants in DIA analysis? Ans: None. The contaminants referenced in the tutorial were added to the FASTA for DDA search purposes only, and would not be used in DIA analysis. Still, they are likely part of the background proteome of the experiment. Giving Skyline this information might allow it to identify issues like peptides occurring both in the contaminants and some other protein of interest, which you decide to target. Q: As far as I know, Skyline keeps only the best spectrum for each peptide to build a spectral library. In that case, the relative product ion intensity might be less robust. Why doesn't Skyline consider variability among fairly ʺgoodʺ multiple spectra for each peptide? Ans: Indeed, we are coming to agree with this viewpoint for this application, and the library build form may soon contain an option to use multiple high quality spectra to create an averaged spectrum. One investigation from ISB found that 5-6 spectra were an improvement over 1 but that improvement beyond that was very limited. As stated above, we believe that a chromatogram library, based on chromatogram peak areas extracted from DIA data, will achieve even greater consistency with future DIA peak area measurements on the same instrument. Our current best-spectrum algorithm also performed better than a straight consensus on all spectra collected for PRM data, where we see many more low intensity, high-noise spectra collected that are still identified as a peptide with high confidence. The relatively recent realization that we may be able to average only a small number of "best spectra" to achieve the majority of reduction in variability may be the key to solving both problems in one implementation. Q: Is there a size limit for the amount of raw data handled by Skyline? Ans: There is not really a limit to the amount of raw data Skyline can handle. Certainly it can target peptides in hundreds of runs both for DIA data, where this may represent hundreds of Gigabytes of raw data, as it can target peptides in hundreds of runs of SRM data. The key factor in memory consumption with Skyline is the number of chromatograms times the number of points in each chromatograms times the number of replicates, or the sum of all chromatogram points stored. 100 peptides over 5 minutes in 100 runs will perform very similarly in Skyline whether the source is DIA or SRM, after the initial import. Initial imports will be much faster with SRM data, since vendor software can retrieve SRM chromatograms much quicker than it can supply all of the points in the mass spectra for a DIA run. Q: I am interested in some glycopeptides that were identified manually in the DDA data. Is there a way to add these spectra into a Skyline library? Ans: Yes, you can use the SSL format described on the BiblioSpec library building web page (https://skyline.gs.washington.edu/labkey/wiki/home/software/BiblioSpec/page.view?name=BiblioSpec%20input%20and%20output%20file%20formats). This is a tabular format you could create in Excel, or write a simple tool to export. You can use it to refer to .ms2 files, which are text representation of MS/MS spectra. Q: Have you guys compared Skyline performance on centroid- vs profile-acquired MS2 data? I'd guess that Skyline is faster for centroid data. However, how about the accuracy of peak detection? Ans: Some comparison has been done, and certainly extracting chromatograms from centroid data is quicker. The primary bottleneck in processing full-scan mass spectra is getting the spectrum m/z, intensity pairs from the vendor software responsible for reading the raw data. The centroided spectra being smaller means the vendor software returns them much more quickly. Skyline's own chromatogram extraction algorithm using binary searching to avoid its ever becoming the performance bottleneck in processing even extremely dense profile spectra. Using centroid spectra, however, obscures a lot of information about the underlying data. It will help with chromatogram peak picking, because it has the potential to remove interference outside your peptide peak of interest, but it is not able to improve upon your instrument's resolving power when dealing with interference inside the time range you are integrating for your peptide. There is a nice explanation of this in a DIA tutorial currently in preparation to be released to the Skyline web site. Q: What about the possibility to refine good transitions if some of the ones that were selected from DDA were found to have some interference? And that would end with some peptides having less than three good transitions for quantification. Ans: Absolutely, DIA data has the advantage that you can refine the product ions extracted over and over until you are satisfied without needing to re-acquire the data. You can start out extracting chromatogram for more transitions than you need, and then use the Edit > Refine > Advanced form to remove all but the best responding transitions. Q: Can I use a text file or excel document (containing protein name, peptide sequence, precursor mz, product ion mz, relative intensities, Rt, etc.) to generate a spectral library in Skyline? Ans: Yes, Skyline supports files of this description called "Assay Libraries" which are commonly used with other tools that perform DIA data processing. Read this small tutorial / tip for more details: https://skyline.gs.washington.edu/labkey/_webdav/home/software/Skyline/%40files/tutorials/ImportingAssayLibraries-2_6.pdf Q: The first of the deleted transition was shown green -- that meant "good" in Skyline. But is the only possible way to eliminate it, to delete it manually? Ans: The green dot means only that it was accepted by Skyline as part of the selected peak group. This may be the case for a lot of poor quality transition chromatograms, if the Settings > Integrate All option is checked, since that forces Skyline to accept all transitions that have any signal at all. With this option off, it is possible to use Edit > Refine > Advanced to eliminate transitions that were not included in a peak group. However, even with the option off, the Skyline algorithm for inclusion in a peak group is not very sophisticated, and may miss many poor quality chromatogram peaks. We are discussing was to improve the automated detection of poor quality transitions in DIA data. For now, though, you will still have to rely on manual inspection and deletion in some cases. Q: During the DIA analysis, we do not collect the full scan data -- does that matter for the DIA quantification? or we can just use the fragment ion for quantification? Ans: On average, you will achieve better dynamic range using fragment ion quantification with a modern high resolution instrument, and narrow enough precursor isolation windows. The Aebersold lab has shown this to be true of the 25 m/z windows they proposed for SWATH. However, we are also working on a manuscript with the Gibson lab, showing the value of including precursor chromatograms extracted from MS1 scans in your DIA data. Whether or not you use them for quantification, they do provide important information for identity validation. Especially for DIA with wider isolation windows, it is possible to isolate modified forms of a peptide (e.g. Methionine oxidation) that share many of the same fragment ions with the unmodified form, making the peaks for the two very difficult to distinguish without precursor information. It is relatively low impact on your data acquisition to include MS1 scans, and that extra information is proving useful. If you include it MS1 scans in your "digital record" then you can always query that information with Skyline. If you do not, then you will not be able to recover it after the fact. Q: How would this workflow differ if we performed PRM instead of DIA on a similar platform? Proper retention time scheduling would likely be an issue, right? Ans: Yes, PRM for more than a handful of peptides would require scheduling. PRM also requires creating a targeted method before starting acquisition, and you can't go back to the data and query new targets after the fact. On the positive side, PRM is much more selective than DIA, and in some cases even SRM. The data files will likely be smaller, and they can be searched with a standard peptide spectrum matching tool, like Mascot or SEQUEST, because the isolation windows are usually very similar to those used in DDA, and so the resulting spectra can be matched to peptides very well, which can add confidence in correct peak assignment, similar to what we have shown for MS1 chromatogram extraction from DDA data, but using more selective product ion chromatogram extraction from MS/MS spectra. Q: So far, what's the max number of DIA runs have been analyzed by Skyline for one particular study? Ans: Hard to say for sure, but we have seen multiple data sets that approach or slightly exceed 100 runs. Again, the limiting factor in size is not number of replicates alone, but number of replicates times number of chromatograms times points in each chromatogram. These experiments with around 100 runs were for hundreds to around 1000 peptides over perhaps 10 minute extraction windows. Q: How would you do internal standard normalization with DIA data in Skyline? Ans: Skyline supports assigning certain peptides in your analysis as "global normalization" standards. You could certainly use that method. In discovery experiments global normalization is frequently performed by normalizing to a median signal or a total of everything measured. With targeted data processing, however, you cannot necessarily expect that the majority of the peptides monitored will not be changing in some interesting way, so that normalizing by a total or a median may actually normalize away the effect you are attempting to measure. It is also important to note that using DIA does not preclude using stable isotope dilution reference peptide or protein standards, which will give you much more precise normalization. If your target list gets large enough, it may still be beneficial to use DIA for the benefit of not needing to worry about scheduling. A grad student working in the MacCoss lab on measuring 700 transcription factor proteins became a big fan of DIA, despite the fact that he had synthetic proteins to normalize against. Q: Why do you choose ʺautoʺ for the max Ion Time in DIA experiment for the MSMS? Ans: We use the "auto" setting for the maximum inject time on the Q-Exactive DIA scans to ensure that our scan rate does not dip far below our expected 10 Hz acquisition rate. The Orbitrap in the Q-Exactive takes 64 milliseconds for mass analysis at resolution 17,500. As the Orbitrap performs this mass analysis, ions are accumulated in preparation for the next MS/MS scan. If the ion accumulation takes longer than 64 milliseconds*, the Orbitrap will sit idle as it waits for ion collection to finish. This idle time leads to a scan rate lower than the maximum acquisition rate of 10Hz. We can't allow the scan rate to dip much below 10Hz in this method, because that would result in a slower duty cycle and potentially having too few points across our chromatographic peaks for accurate quantification. In our hands, we find that the "auto" setting allows a long enough injection time for the majority of scans to hit their AGC target. However, some experiments may require longer inject times to collect enough ions. In these cases, the number of scans per duty cycle should be reduced by reducing the mass range to be covered in each injection, or increasing the isolation window width of the MS/MS scans.
Q: Any thoughts on DIA for phosphopeptide data? Poor fragmentation, site localization, etc. Ans: If you use the workflow described in this tutorial, then your site localization will be only as good as your DDA peptide spectrum matching software. However, Skyline will expose the retention time dimension of your data to you in a way that is not often as easy with spectrum matching software. In looking at phosphopeptide data this has proven extremely useful as different phospho site localizations usually differ in retention time, causing multiple peaks in chromatograms. This can help to identify when a spectrum matching tool has made a mistake, like assigning multiple spectra acquired during the same chromatogram elution curve to different phospho site localizations. Also, through more careful preparation, using the "Prior Knowledge Workflow" described in this session, you can build a library with accurate fragment ion relative ion abundance and retention time information for various phosphosite localizations and use that to query your DIA data. Including MS1 scans in your DIA acquisition schemes may help you to work around poor fragmentation, as it allows you to fall back to quantification using the MS1 chromatograms, or essentially what has been described as accurate mass and time (AMT) analysis of MS1 data. Much of modern DIA data analysis is essentially AMT on product ion chromatograms extracted from MS/MS. Q: Can Skyline process Waters MSE data? -- since it's also a kind of DIA analysis? Ans: Yes, definitely. There is a special "Isolation Scheme" specifically for this, called "All Ions". Select it, and Skyline will apply MSe logic for Waters data (alternating low and high energy spectra). Skyline also supports building spectral libraries from PLGS final_fragment.csv files so that you can use peptide identifications in your imported data files, similar to what was described for DDA data in webinar 1. And, Skyline now supports chromatogram filtering in the ion mobility drift time dimension for HDMSe data. Without this dimension, however, we recommend that you use the precursor chromatograms (extracted from low energy spectra) for quantification in standard MSe data, since the lack of any isolation for the high energy scans reduces the selectivity to the point where it becomes less selective than MS1 (low energy) data. We work closely with Waters on supporting their DIA/MSe data analysis requirements in Skyline. Q: MSX processing is very slow, will it be better in the future? Ans: I apologize for this, I do realize that the import of MSX data is slow. We have plans and ideas for improving both the quality and the speed of MSX demultiplexing. We are also developing other multiplexed alternatives to MSX that process more quickly. Unfortunately, I can't make any promises as to when these approaches will be ready for Skyline implementation as the implementation is far from trivial. Q: What is the maximum retention time variability allowed for succ. quantification? Ans: There is not a simple answer for this. Chromatography-based quantification, however, does depend on your chromatography, and the better your chromatography in RT consistency, peak resolution and spray stability the better your quantitative results will be. We highly recommend implementing quality control metrics that you run frequently among all your mass spec runs to measure your chromatography and catch issues early. Good places to start is by looking at SProCoP: Q: Using skyline to processing MSX big data: It is impossible to handle-by-handle -- is there a way to evaluate big data? Ans: I'm assuming the question meant to say "handle by hand". I agree, for large experiments, where many peptides are being analyzed, and there are many replicates, manual analysis of MSX, or any DIA data, is too time consuming to be practical. We are very focused on improving automated analysis of DIA data. We found that the most difficult aspect of automated analysis of DIA data was picking peaks in the data, and have made great strides in that area. If you haven't already, I highly recommend trying out some of the advanced peak picking models in Skyline. We have a tutorial for that here. Another issue will be transition refinement. In this case, it really helps to have some prior knowledge, such as DDA data. We are focused on improving this aspect as well. Q: Can you use Skyline to determine data ʺdensityʺ and hence design optimal isolation windows with variable sizes (i.e. smaller windows in mass range with more peptides eluting)? Ans: Not sure how Skyline would be used to determine data density, but such models have been derived (including in our lab) and could be derived again elsewhere. Obviously, you would need to derive your data density expectations before creating your DIA method and acquiring your quantitative data, making this something of a targeted method. In practice, however, data density models seem to be derived relatively infrequently and the expected to apply to many sample matrices. In fact, the 400-800 range used in this webinar and its more common variant in our lab 500-900 are based on a data density model derived by Jarrett, based on peptide ID counts in DDA data. The range 500-900 was determined to capture the vast majority of peptides of interest. The Skyline isolation scheme editor is certainly capable of accommodating varying isolation windows, and was, in fact, designed with this in mind. Q: Is Skyline able to estimate precursor number in different time points of elution to make a non-similar SWATH window distribution across the gradient? To use narrower windows for more rich in peptide precursor rtime regions? Ans: Skyline does not yet explicitly support "scheduled DIA" as you seem to be describing it. To our knowledge, this is not well supported by the instrument control software either. Skyline does support varying windows across the m/z dimension, but not yet the RT dimension. It seems likely that doing the latter would require new collaborative efforts with the instrument vendors, and new experimental efforts to prove that varying isolation window widths over the elution of a peptide would produce valid quantitative results. Q: Can you process Fusion WiSIMDIA data? Ans: Apparently there is a bug in the current processing of WiSIMDIA data in Skyline. We will fix this soon. Q: If I didn't identify my target proteins from the DDA run used to build the library, is it possible to find the peptide in the DIA run by listing it in the target window? Ans: It is possible to target peptides you did not find in a DDA run interspersed with your DIA. The workflow presented in this webinar is just the simplest way, we believe, to get started with DIA data analysis. The workflow described as the "Prior Knowledge Workflow" for DIA data analysis, will allow you to query any peptide from DIA with high confidence, providing you start with prior knowdge of retention time and fragment ion relative abundance of the peptide, which can be derived from many sources other than a DDA run on the same sample matrix you use in your DIA runs. e.g. targeted runs on synthetic peptides or proteins, simplifying sample processing like fractionation, etc. Once you have this prior knowledge, you can very effectively query your DIA data for peptides you didn't find with DDA runs. Q: Are there plans to fix the MS1 extraction from WiSIM data on the Fusion? Right now Skyline tries to extract precursors from SIM scans that do not actually contain the right m/z value Ans: Yes, we will fix this soon. Q: Compared to QE, how does the DIA setup differ on Fusion? Any advantages on Fusion? Ans: The set up for DIA is a little bit different on the Fusion. For the fusion method, you'll have 4 experiments total: 1) MS1 2) tMS2 3) MS1 4) tMS2. The MS1 scans will be the same as on the QE. The first tMS2 experiment will have the first ten isolation windows in the DIA list output by Skyline. The second tMS2 will have the second half of the list. On the Fusion, the isolation window targets have to be split out into separate lists because there is no "loop count" setting. The tMS2 scan will always go through its complete target list before the instrument moves onto the next experiment. If you were OK with an MS1 scan happening every 20 MS/MS scans (I often do this), you can have just a single MS1 experiment, and a tMS2 experiment containing the entire inclusion list exported by Skyline. Also, keep in mind that you'll need to remember to specify HCD fragmentation and Orbitrap mass analysis on the Fusion for the tMS2 scans. As far as advantages, the Fusion is capable of performing DIA in a very similar manner to the QE because it has very similar components (quadrupole, HCD cell, Orbitrap) to the QE. One benefit of the Fusion hardware could be trying out DIA using resonance CID for dissociation in the linear ion trap, which the QE can't do, but I haven't looked into that extensively. Q: Could you please comment if you use FDR for your DIA data, and if yes - how do you generate decoys? Ans: Yes, the full details of performing integration of DIA data in Skyline that makes use of an advanced peak picking model and generates q values are covered in the Advanced Peak Picking Models tutorial. We now feel that the "Shuffle Sequence" method of decoy generation performs moderately better than "Reverse Sequence" or just using the second best peaks. We do not recommend using the "Random Mass Shift" method described in the original mProphet paper for DIA data. All that said, we should also note that these calculated FDR values are for peak detection only, and not for quantitative values. Good quantification usually demands a lot more attention to detail than peptide spectrum matching, and a PSM or peak detection FDR of 1% does not immediately translate to 1% quantitative FDR. There are many considerations like correct integration boundaries, interference, phenomena like peptide degradation and potential confounding from system changes over time when sample processing is not adequately randomized and blocked that will impact quantitative accuracy without being captured in peak detection FDR values. Q: Is the DIA data analysis done in the same way you presented if the method is prepared with multiplexed m/z windows (in a QExactive)? I am getting started with DIA and i would try to do a 5x4m/z method. Ans: There are some differences in setting up the instrument method. When calculating the Isolation scheme (under Transitions Settings --> Full Scan), you will want to select a window width of 4 m/z instead of 20 m/z, and be sure to select "MSX" under deconvolution and enter 5 windows per scan. Then, in Transition Settings --> Instrument, you'll enter 5000 for the Firmware inclusion limit. This is necessary because MSX outputs a very long inclusion list for the instrument to loop through. Because of the length of the list, we need to be sure to not output more isolation window than the instrument can handle. That's all you have to change on the Skyline side of things. But you will also have to set up your instrument method correctly. For the QE, you'll need to turn on MSX, set it to 5, and I recommend a maximum injection time of 20 ms (per window). I have a tip up on the Skyline web page that discusses this here. How would you do internal standard normalization with DIA data? Ans: As mentioned above, you could use the ability to mark certain peptides in your Targets list as "Global Normalization" standards, if your experiment is label free. This is the best option for a truly targeted label-free experiment, since you should not rely on the average targeted peptide to be representative of system performance for a peptide expected not to change, an expectation frequently employed with discovery data when quantile normalization is used. There is, of course, nothing about the DIA method that precludes its use in labeled experiments. In large labeling experiments DIA provides benefits over SRM and PRM, which may require multiple runs to measure all peptides of interest, and even in relatively small experiments, DIA alleviates the need for scheduled acquistion. Skyline provides extensive support for normalization by labeled reference peptides. Q: What is the maximum mass of a peptide that can be analyzed by DIA? Ans: The maximum mass will depend on your instrument and isolation scheme and the charge state of the peptide, since the instrument and isolation scheme will define a maximum m/z. With a high enough charge state, even fairly large peptides will achieve m/z within the normal maxima of DIA experiments (e.g. 900 or 1200). As long as your instrument will support it, you can extend your isolation scheme higher, but as presented in the webinar, you should be considering your overall "image quality" for your DIA isolation scheme. Extending the range will come at a cost, either by moving the minimum up with the maximum, or in selectivity or cycle time. These forces are competing for your overall DIA image quality. Q: Can we import data from Peaks search engine? Ans: Yes. The full list, to the best of our knowledge, of supported spectrum matching pipelines is listed here: Q: What are the numbers in paranthesis next to each peptide Ans: Those are the indices of the starting and ending amino acids for the peptide in the protein sequence. Unfortunately, the indices are zero-based, because Skyline was designed by a programmer, and not 1-based as is more common in the field. This may get fixed some day, but it has been true for over 5 years now, and mostly people seem to have adjusted to it. Q: Is Skyline able to estimate precursor number in different time points of elution to make a non-similar SWATH window distribution across the gradient? To use narrower windows for more rich in peptide precursor rtime regions? Ans: As explained above, Skyline does not provide help with estimating peptide density either in the m/z or RT dimensions. Skyline does support specifying varying isolation windows in the m/z dimension to account for such density differences. It does not yet support varying windows over the RT dimension. Nor have we heard of an instrument setup that supports this or an experiment that proves such scheduled DIA acquisition produces valid quantitative data. Q: How important is Percolation of PSMs for then getting a good library and ultimately, quantitation? Ans: The Skyline library builder requires some kind of probability value to use for a cut-off. The probability value is achieved in many different ways in different spectrum matching software. The only ones that use Percolator, to our knowledge, are Mascot and SEQUEST. Q:When adding peptides from a spectral library and associating the proteins, why does it match semitryptic peptides with non-specific cleavage at the C-terminus, but not those at the N-terminus? This has implications for association of N-terminal peptides after signal peptide cleavage Ans: This is an issue that has come up recently at the MacCoss lab. We are thinking about it, and we hope to have it fixed for the next release. Thanks for helping to increase the priority on this issue: https://skyline.gs.washington.edu/labkey/issues/home/issues/details.view?issueId=365 Q: Can OpenSWATH be used along with Skyline for data generated from AB Sciex Triple TOF instrument? Ans: Yes. We have worked closely with the creator of OpenSWATH (Hannes Roest of the Aebersold lab), and we have created multiple points of integration, both in allowing the import of integration boundaries from external software and allowing the import of "Assay Libraries", which OpenSWATH also uses. The following tutorial tips provide documentation of this integration: https://skyline.gs.washington.edu/labkey/wiki/home/software/Skyline/page.view?name=other_tools However, OpenSWATH, is somewhat difficult to implement for any lab, as it is designed to run on a Linux cluster. The implementation in Skyline includes many aspects of the mProphet and OpenSWATH implementations, and our own tests show that Skyline now achieve similar performance to OpenSWATH (manuscript in production). For complete documentation on using advanced mProphet-base scoring models in Skyline consult the Advanced Peak Picking Models tutorial: Q: Can you extract a P-value and cut-off for significant peptides/proteins? Ans: As explained above, you can use Advanced Peak Picking Models as described in the tutorial: To calculate peak detection q values (a probability measure associated with FDR, adjusting for multiple hypothesis testing). These values should not be taken as accurately capturing error rates in quantification, but only for whether the correct peak was chosen for the peptide. Once you have calculated these q values, you can use the Document Grid in Skyline to sort by the q value annotation, and then paste a list of peptides above your desired cut-off into the Edit > Refine > Accept Peptides form. Q: Could I merge two peptide's XIC in one graph? Ans: You can show multiple peptide XICs in the chromatogram graphs in Skyline either by selecting a protein, where the chromatograms for all its peptides will be shown in a single graph, or by using shift-selection, click-and-shift-click or control-click to create a multiple selection. Skyline will show chromatograms for all peptides in a multiple selection. For more complete details on this feature consult the following tip: |