Analytical performance
The analytical performance of the ProQuant™ platform comes from a combination of multiple optimisations in the bottom-up proteomics process, including sample preparation techniques, MS optimisations, and data analytics.
One of the key changes is to the way the mass spectrometer collects data on peptides coming off the chromatography column. A typical sample contains hundreds of thousands of peptides, many of which may elute from the column together. In traditional data-dependent acquisition (DDA) mode the MS needs to both collect the information on the abundance and mass of all of the peptides eluting together (the MS1 survey scan) and then select the most abundant MS1 ions for fragmentation (MS2 scans), to provide the data to identify the peptide. Even the fastest mass spectrometer has to make compromises in the way it does this. ProQuant™ manages this compromise much better than our competitors – it’s complex and proprietary, but fortunately all our clients see is better data!
Other important gains come from optimised sample and MS data processing to generate the most reproducible output data possible. We also use machine learning to improve peptide identification and reduce artefacts. Importantly every single one of these processes is unsupervised, completely eliminating any possibility of introducing bias into the data related to the biological question.
The proof is in the data
We have repeatedly tested and proven the quantitative performance of ProQuant™ by comparing our platform with competitor label-free platforms using a variety of bottom-up DDA proteomics methods. We have done this by analysing replicate sets of serum samples and then calculating the analytical coefficient of variation (a measure of the quantitative reproducibility of the method) across all the different peptides present.
In a complex protein mixture ProQuant™ has an analytical CV including sample preparation below 20% for just under half of all of the peptides identified in the samples. The number of peptides quantified at this level of analytical performance was 10x greater than the average of three competitor approaches. Even including our proprietary MS method into competitor approaches (where we are now comparing analysis of precisely the same raw output files from the MS; * in graphs), we still identify 7x more peptides than the average of three competitors at this level of precision, reflecting the quality of our data analytics.
The same analysis at the protein level shows a similar picture. One would expect errors in the individual estimates of the peptides to cancel out somewhat when the abundances from the peptides are combined to generate one estimate for the “total” level of that protein, but ProQuant™ still identifies 2x – 10x more proteins with a CV <20%. This results from a combination of a better estimate of the underlying peptide abundances and the ProQuant™ data analytics improving peptide identification.
What does this improved precision mean for your experiment? By using conventional power analyses (Cohen, 1998) we can calculate the minimum effect size that each of the different methods can distinguish using a given sample size or determine the group size needed to detect a given effect size.
In the table below we show the results of a power analysis in which we have calculated the group size needed to detect a 30% effect size in a simple two-group experiment for 100 proteins from complex samples (serum following depletion of the most abundant proteins).
This shows that using ProQuant™ a simple triplicate will enable you to detect a 30% effect size on 100 of the proteins in your sample, whereas competitor methods require the use of at least four times the number of samples to achieve the same outcome.
Put simply, competitor methods are fine to provide a list of what was detected in the sample but inadequate for estimating the relative amounts of each.
So what about data-independent acquisition? The world never stands still, and other approaches have been developed to try and address the limitations of bottom-up proteomics, of which perhaps the best known is a data-independent acquisition technique (DIA) called SWATH (sequential window acquisition of all theoretical fragment-ion spectra mass spectrometry). But again, we have performed a head-to-head comparison of replicate samples using DIA-SWATH performed by a leading proponent of that approach and ProQuant™ – and once more, ProQuant™ came out far ahead, with approximately 5x more peptides identified with a CV <20%.
Services
We have spent years developing the ProQuant™ platform, which has given us considerable experience in adapting the platform to a wide range of applications.
- Target identification
- Label characterisation
- Biomarker identification
- Biologics CMC
- And more!
See our services for more information.
Case Studies
ProQuant™ has been used to
- Find PTMs in complex samples
- Characterise protein labelling
- Investigate proteoform diversity
- Identify biomarkers
For more details on these applications see our case studies.
Get in touch
Call Us 01223 839557
Email Us [email protected]
Visit Us
RxCelerate Ltd
Dorothy Hodgkin Building,
Babraham Research Campus,
Babraham, Cambridge CB22 3FH