OPINION | Dr Jarrod Sandow, Chief Product Officer

For as long as I have worked in proteomics there has been an assumption underpinning how we design our workflows. If you want the best possible performance, you use direct injection. If you want robustness, scalability, and protection of your analytical column, you introduce a trap. However, the benefits of introducing a trap were always offset by the reduction in performance that you would observe. It’s a classic frenemy scenario.

Trap columns exist for good reason (my tendency to generate chunky samples in my early postdoc years is one example). They concentrate sample injections, enable rapid sample loading, and provide on-line clean-up that protects analytical columns from contamination and blockages (and early-career postdocs). With an increased focus in the field on high-throughput sample analysis, traps have also become an essential component to driving workflows faster. Yet despite all this, the prevailing mantra is traps equal reduced identifications, reduced sensitivity and reduced proteome coverage. If performance mattered most you avoided traps. If robustness mattered most you accepted a drop in performance which, in some cases, amounted to a reduction of more than 10% in identifications.

In developing the NanoShield® C18 trap column, our objective was to pair the chemistry and performance of the trap perfectly with our range of analytical columns to eliminate compromise. Across multiple LC–MS platforms and gradient lengths, we observed less than 1% reduction in protein identifications and only a 1 to 3% reduction in precursor identifications compared to direct injection. Peak quality metrics, including FWHM and peak shape, remained effectively unchanged. These results are also consistent across a range of different applications, including challenging workflows such as single-cell proteomics. We were very excited as we’d eliminated the typical compromises associated with the introduction of a trap while retaining all the benefits. The historical performance gap between trap-based workflows and direct injection had effectively closed.

In addition to compromising chromatographic performance, traps that were poorly matched with the analytical column would often fail to bind subsets of peptides during loading, sending peptides to waste rather than into your mass spec. One of the more persistent challenges with traditional trap columns has been the loss of hydrophilic peptides. They are often central to the biological questions researchers are trying to answer particularly in applications rich in these peptides such as glycoproteomics, phosphoprotemics and labelled peptide quantification. With NanoShield®, we set out to ensure that the trap and analytical column chemistries were perfectly paired to eliminate the loss of peptides to waste. When we compared the NanoShield® results to an alternative trap that is common in the field, we saw a significant increase in hydrophilic peptide retention at the start of the sample gradient. For me, this is one of the most important aspects of the technology. It doesn’t matter if you have narrow and symmetrical peaks if you’re missing large subsets of your sample in the analysis. When these peptides are lost during sample loading you don’t only see a reduction in identifications but a distortion of the dataset itself.

With an increased focus in recent years on high-throughput analysis of samples, the benefits of using a trap have extended beyond just sample clean-up and concentration, increased robustness and protection from an early-career postdoc. The ability to load samples onto traps faster than an analytical column allows a greater proportion of a sample run to be dedicated to sample separation instead of sample loading. This reduction in overhead times effectively lengthens the sample separation time, improving separation efficiency and the utilisation of mass spec time for sample analysis for each run. As the field approaches hundreds of samples per day becoming the routine approach to proteomic sample analysis rather than the exception, these gains in efficiency and better utilisation of mass spec time are essential elements of new LC-MS workflows. 

What excites me most about the NanoShield® is that we can finally move past traps being a “just in case” addition to an LC-MS workflow and accept that this would come with compromises. We accepted that certain trade-offs were unavoidable, and optimised within those boundaries. But when those limitations are removed we can begin to utilise traps to improve our workflows and not just to protect them. We no longer have to choose between sensitivity and robustness. We no longer have to protect our systems at the expense of our data. We no longer have to accept incomplete coverage as the cost of operational stability. Instead, we can begin to design workflows that are both high-performing and resilient from the outset.

Proteomics is evolving rapidly, and as we push toward deeper coverage, higher throughput, and more demanding applications, the expectations placed on our workflows will only increase. Meeting those expectations requires not just incremental improvements, but a willingness to turn a former frenemy into our best friend.