NEW YORK (GenomeWeb News) – As mass spec-based proteomics continues to move into the clinic, questions of accuracy and reproducibility have become increasingly significant.
With this in mind, researchers from Leiden University Medical Center performed an analysis of where error and imprecision arise in typical MRM mass spec workflows, looking at assays for quantifying serum apolipoproteins Apo A-1 and Apo B.
Detailed in a paper published last month in the Journal of Proteome Research, the LUMC study identified trypsin digestion as a key point in MRM workflows, finding that while digestion added little to assay imprecision, it was a considerable source of inaccuracy, Irene van den Broek, a LUMC researcher and first author on the paper, told GenomeWeb.
This was somewhat surprising, she noted, as she and her colleagues were initially concerned that digestion would be a major cause of imprecision.
"However, that was not the case," she said. "We could get surprisingly good precision – between 5 and 10 percent [CVs] for all the peptides that we tested – and we didn't see any reduction in imprecision when we tested digestions of four hours or 20 hours."
In fact, van den Broek noted, the primary cause of imprecision in the team's assay was liquid handling prior to addition of standards – an issue that, she said, can be significantly improved by automating these steps, which the researchers have since done.
The accuracy of their measurements, however, was another story. Here, van den Broek said, differences in digestion efficiency played a significant role, leading to challenges in arriving at measurements that were not only reproducible but also accurate.
This is a problem inherent in targeted proteomic workflows wherein peptide levels are used as proxies for whole protein levels. Because the measured peptides are created by digesting the protein of interest prior to mass spec analysis, differences in digestion efficiency that lead to differences in the amounts of peptides released can affect an assay's accuracy.
Van den Broek and her colleagues observed such differences in their apolipoprotein analyses, finding, for instance, that they were unable to generate truly linear calibration curves.
"We knew the exact protein concentration, and still the linearity wasn't good enough," she said. Looking at the digestion efficiency in roughly 100 samples, the researchers also found variations in digestion efficiency that could lead to inaccurate measurements.
While this digestion efficiency issue didn't lead to significant inaccuracy in the majority of samples, it does present a problem for clinical assays, van den Broek said.
For instance, "in one of our calibration samples we had a bias of minus-10 percent for Aop A-1, and we were aiming for a 5.3 percent bias, which was the minimal requirement for bias based on biological variation," she said. "So that would have been significant [clinically]."
Most samples had bias within this 5 percent, but if there is just one in 100 that has this bias of 10 percent, that is not good," she added.
The findings, van den Broek said, support the use of full-length stable isotope labeled proteins as standards as opposed to SIL peptides. SIL peptides are less expensive and easier to obtain than full-length proteins, and so are more typically used as standards. But, said van den Broek, while both types of standard appeared to work equally well in controlling for imprecision, use of the SIL proteins allowed the researchers to detect bias resulting from differences in digestion efficiency that would have gone undetected using the SIL peptides.
"I think for very precise measurements you don't need SIL proteins," she said. "But if you want to compare your results with other labs, then it comes down to the trueness [of the measurement], the bias, and I don't know if we are at the level yet to compare based on peptide standards. That is really only possible if you know that you are really at 100 percent digestion recovery with no loss at any point."
This is particularly important for porting assays across labs, van den Broek said, noting that different reagents and set-ups in different labs can result in different levels of peptide recovery, "so you need the calibration [with SIL proteins] in place so that you [arrive at] the same protein concentration."
She and her colleagues are now looking into various ways to improve digestion efficiency, including use of denaturing reagents to better open up proteins for digestion and use of reagents like immobilized trypsin.
Immobilized trypsin "could be a huge improvement" in that in addition to improving efficiency, it offers faster digestion and is easily compatible with automation, van den Broek said. "I think for real clinical adoption [you need] a method that is fast and easily automated and also has a very consistent digestion efficiency."
Immobilized trypsin also has the advantage of modifying trypsin to prevent autolysis, which can lead to chymotrypsin-like activity that causes peptide decay during digestion and can lead to additional bias.
For instance, a 2012 study led by North Carolina State University researcher David Muddiman found that this phenomenon led to greater than 30-fold differences in protein measurements in an investigation of quantitative mass spec assays to 24 enzymes.
As Philip Loziuk, a graduate student in Muddiman's lab, told GenomeWeb in an interview last year, "we realized that [target] peptides were decaying over the course of [the 16-hour] digestion, and the problem was they were being produced and decaying at different rates," which led, in some cases, to quantitative error of greater than 100 percent.
In a follow-up study published in 2013 in the Journal of Proteome Research, Muddiman's lab found that using modified trypsin – like in an immobilized form – eliminated the peptide decay observed in the MCP study along with the associated bias.