Part of a heatmap at Review Alignment
When we released Progenesis LC-MS v4.1, one of its key improvements was the addition of quality control (QC) measures to the retention time alignment. If you’ve used Progenesis before, you’ll probably see them first in the heatmap shown at Review Alignment. (If you’re new to Progenesis, it may be worth taking a minute or two to read the Alignment overview first.)
However, before we dive into the finer details of the new tools and visualisations, let’s take a look at the basis for the improvements we’ve made…
Why was this introduced?
Although the set of tools available to support it is still quite limited, quality control is a massively important aspect of any experiment. Indeed, Progenesis software has always acknowledged the importance of QC (e.g. Progenesis SpotCheck for 2D PAGE analysis), providing excellent visualisation tools to help validate your results, rather than merely presenting you with an answer and expecting you to believe it.
While the matrix of views and editing tools we present in Alignment is one such example of this ability to review, we realised that alignment reviews could be both time-consuming and open-ended; a common question we heard was “How do I know when my alignment is good enough?”. Given the degree of variation between proteomics experiments, this was an extraordinarily difficult question to answer!
What we needed was an objective measure of the alignment quality that could be used by researchers like yourself, to guide your review and provide greater confidence in your analysis.
How have we improved matters?
In Progenesis LC-MS v4.1, there are now measures that indicate the quality of your experiment’s retention time alignment. This is reported at two levels:
- Each run has an overall alignment quality score, shown in a new column in the list of runs at the left; clicking the column header to sort by these values can help to find any runs that have low values. Although it varies from experiment to experiment, values over about 80% can typically be considered good.
- Each run’s ion map is also split into a 16×16 grid, with each grid square scored individually and given one of three quality ratings – Good, OK, and Needs Review – and a corresponding colour of green, amber or red. (In fact, there’s also a fourth rating – Unknown – which is used when the signal intensity is so low that a reliable conclusion can’t be drawn from the data. In this case, no colour is applied.)
The location of the new alignment quality measures, seen in the Review Alignment step of Progenesis LC-MS v4.1
These run scores and ratings are calculated automatically and can be used to guide your review, as we’ll see later.
How do the alignment ratings work?
The rating for each grid square is based on how well the signal intensities in that area of the alignment reference run match those in the corresponding part of the run being aligned – after applying the alignment, of course. Rather than being based purely on chromatographic alignment, this takes into account the m/z position of ion intensities too, helping to avoid false positives.
The overall alignment score for the run is then determined by taking into account each of the grid square scores and weighting them according to the strength of the signal at that point in the run.
What if I disagree with a poor alignment rating?
In short, if you disagree with the rating, then you can ignore it. The important thing is that you review the alignment – remember that the rating is “Needs Review” and not “Poor”.
Heavy noise can be rated as Needs Review
Occasionally, you may find a region of your run marked as Needs Review, but when you look more closely, you find little evidence for the low score. In our experience, this only arises when an area of the run consists of nothing but a large amount of randomly distributed – but intense – noise. Thankfully, such situations are rare; when we were testing the new QC features (using a wide range of datasets of varying run quality), that certainly proved to be the case. If you find that you often disagree with the rankings, please get in touch, as we may be able to help.
Of course, it’s also good to remember that the scores are only there to help you review and validate your experiment’s alignment. The scores don’t affect the quantitation in any way – they merely offer a measure of confidence in the alignment quality.
In summary: your new alignment review workflow
Putting all of this together, here’s a good strategy for reviewing the results of alignment in your experiments:
- Set up the options to enable simple reviewing of the alignment quality in individual runs:
- Display the alignment quality heatmap by setting the relevant option in the View menu
- Set the focus grid to the same size as the heatmap grid squares: deselect the “Make box square” option and select the “16” focus grid size
- In the list of runs, click the Score column’s heading to sort them by their overall alignment score, and start by selecting the run with the lowest score
- Click on any grid square in the Ion Intensity Map to inspect that area in the Vector Editing and Transition panes (try starting with any “Needs Review” squares)
- If the ions in the grid square are in fact badly aligned, add a manual alignment vector in the Vector Editing pane to improve the alignment
- Repeat steps 3 and 4 for all runs in your experiment, using the heatmaps to guide you
Of course, don’t forget to check some of the good scores and ratings too – it’s just as important to know that those ratings are correct as it is to review (and possibly improve) the alignment of the lower-scoring areas.
Overall, I’m confident that these new features in Progenesis LC-MS will make your alignment review process faster, easier, more objective and more consistent. And, hopefully, the question of “How do I know when my alignment is good enough?” will now be answered. Speaking as someone who has helped others with that very question in the past, I know it’s already helping me. To see for yourself how it can help you too, download Progenesis LC-MS today.