5 key steps to peak picking that make Progenesis QI better

Nonlinear Dynamics were at MetSoc2015 this year, talking to attendees about our small molecule software Progenesis QI. One common confusion we discovered was our use of the term peak picking. In the minds of many, this simply refers to detecting MS peaks in profiled LC-MS data. In Progenesis QI, we use this term a lot more broadly, to refer to the entire process of reducing your data from MS spectra to a set of detected compounds and their quantitative measurements.

In fact, what is commonly termed peak picking is performed in Progenesis QI at the Import Data phase. Here, your raw data is reduced to a set of peak models, which reduces storage requirements significantly, whilst retaining all necessary quantitative information. Importantly, our peak modelling process retains information about the shape of your peaks, which is lost in typical centroiding implementations.

After data import and peak modelling, the next step is to align your runs in the retention time direction. This enables the first step in the peak picking process:

Step 1 – Run aggregation

Progenesis QI employs co-detection, which means that compound ion detection is performed once on a single aggregate run. The advantage of this approach is that you obtain no missing values.

In this first step, Progenesis overlays the runs selected for peak picking to produce a single aggregate run. You may notice that early on in the peak picking process, the ion map at the Peak picking screen updates. This means the aggregation process has finished, and the ion map has been updated to show the aggregate run.

Aggregate run has been created in Progenesis Figure 1. An example section of the ion map of an aggregate run. Compound ion detection is performed once on this single aggregate, and detected outlines are then measured on all runs to ensure a measurement for all compound ions on all runs.

Step 2 – Chromatogram detection

The next step is to transform the MS spectra into a list of chromatograms. In basic terms, peak models at similar m/z’s in neighbouring mass spectra are joined together to form a chromatogram. Progenesis also analyses the shape of the chromatogram to distinguish overlapping chromatograms for partially co-eluting compounds. If your data contains ion mobility information, Progenesis will also use this information to distinguish co-eluting chromatograms.

The chromatograms have been detected on the aggregate run Figure 2. The same section of the aggregate run, with the detected chromatograms highlighted. Chromatograms are formed by joining together MS peaks from neighbouring spectra, where the m/z of the peaks coincide.

Step 3 – Isotope deconvolution

Very often, multiple isotopic forms of the same compound ion will be detected in your data. Progenesis analyses the detected chromatograms and groups together chromatograms which appear to be isotopes of the same compound ion. If more than one isotope of an ion is detected, Progenesis can also determine the charge of the ion, by analysing the m/z shift between isotopes.

Image showing isotope deconvolution on the ion map Figure 3. The same section of the aggregate run, showing the result of isotope deconvolution. In this case, the three chromatograms have been joined into a charge 1 compound ion, since the isotopes are all ~1 Da apart.

Step 4 – Adduct deconvolution

One often detects multiple adducted forms of the same compound or peptide. Progenesis QI will analyse the retention time profiles and m/z separations of your detected ions to group compound ions together into compounds. If multiple adducts of a compound are detected, Progenesis will also calculate and show the neutral mass of the compound. You can manually review the results of this process at the Review Deconvolution screen.

In QI for proteomics, a slightly different but analogous process is performed when peptide identifications are imported. For each identified peptide ion, its identifications are propagated to all peptide ions that deconvolve with it. This means that if your database search only identifies one charge state of a peptide, all other charge states will be given the same identification by Progenesis.

Adduct deconvolution in Progenesis QI Figure 4. The result of adduct deconvolution for the compound ion shown in previous figures. Here it has been determined to be an M+H ion, and two other adducted forms of the same compound have been grouped with it. The compound has been assigned a neutral mass of 472.2542 Da, based on the masses of the adducted forms.

Step 5 – Quantitation and normalisation

Accurate quantitation information can then be determined for each of the detected compounds. Note that because we perform isotope and adduct deconvolution, the quantitative measurements more accurately reflect the total amount of a given compound present in your samples. Since detection was performed on the aggregate run, every compound and compound ion can be measured on every run (this is the no missing values approach referred to earlier).

Finally, Progenesis normalises the quantitative data across all runs in your experiment. This corrects for experimental or technical variation when running samples, and allows comparison of abundances between conditions in your experiment.

Abundance profile graph Figure 5. The normalised abundances for the compound shown in Figure 4. The compound abundance on a given run is the sum of its constituent ions’ abundances on that run. Co-detection means that every compound can be measured on every run. The abundances are normalised across runs so that comparisons can be made.

Summary

As you can see, the Peak Picking screen in Progenesis QI does much more than you might assume from its name. Progenesis QI employs a comprehensive set of algorithms to reduce your raw data from a list of mass spectra into accurately quantifiable, isotope and adduct deconvoluted compounds.

Strictly speaking, then, the Peak picking screen in Progenesis should be renamed Run aggregation, chromatogram detection, isotope deconvolution, adduct deconvolution, quantitation and normalisation. But we thought that didn’t quite fit into the workflow UI 🙂

If you have any questions or want to find out more about Progenesis QI, please check our FAQs, or get in touch.

Post a Comment

Your email is never shared. Required fields are marked *

*
*