Is this the future of LC-MS/MS?

' Dr. Armel Nicolas

Mass spectrometry-based proteomics is an amazing and still fast evolving field. The way clever people constantly come up with new ways to tweak experiment design to tackle wholly different biological questions never ceases to amaze me. When I started in MS, I had mostly heard of people using MALDI-TOF instruments. Orbitraps were just emerging as a viable mass analyser. Now they are all over the place and ESI has essentially replaced MALDI for most applications.

However, as I learned the basics of the craft, I also realised three of its main limitations:

  • I experienced my first disappointment when I understood that – unlike in genomics or transcriptomics – in MS-based proteomics peptides are not really being “sequenced” (see Database search ). It took me some time to accept that the results from database-searches could be trustworthy. This also meant that one could not do proteomics without a reference search database of expected proteins.

Still from Jurassic Park

While we still don’t have dinosaur DNA, we do have quite a few proteins, so there is an upside!


  • My second great disappointment was to realise that LC-MS/MS only detects what the user allows it to detect (a consequence of the previous point). Peptide variants absent from the search database, or peptides modified in ways not included in the search parameters, will not be detected.
  • My last major disappointment concerned the well know issue of Data Dependent Acquisition (DDA), namely that, because in each duty cycle the MS decides on the fly which precursor ions to fragment, multiple technical replicate runs of a same sample will result in only a partial overlap between lists of peptides identified in each, resulting in a very patchy picture.

Still from the Fly

Not that fly!

But do we not live in glorious times? Lo! All three of these issues may soon be things of the past:

  • While database search engines are still the best and most widespread available tools, good de novo sequencing solutions are becoming available (PEAKS, Novor, pNovo, pepNovo… and there is talk that MaxQuant may soon provide that feature). Of course, de novo sequencing is no panacea, as there is still a high number of peptides with a significant margin of uncertainty; this means that ideally both strategies should be combined.
  • The second limitation is also becoming less and less restrictive: as the community becomes aware of it, and as the number of experiments in which experiment-specific RNA reads are available, tools are becoming available to “update” the expected proteome from RNA, so that sample-specific peptide variants can be detected even when a database search strategy is used.
  • Lastly, issues with Data-Dependent Acquisition… There have been ways to mitigate these issues for a while, e.g. labelling samples, or the Match-Between Runs feature in MaxQuant, both of which do not improve overall coverage but improve inter-sample overlap.

But with the recent advent of Data Independent Acquisition (DIA) methods, things are really changing fast. There are many variant DIA methods: Waters’ version is called MSE, SCIEX offer SWATH (and denied Thermo Scientific the use the name), and Biognosys have HRM (humbly standing for “Hyper”-reaction Monitoring).

All of these variants have in common that they allow generation of mixed, high complexity MS2 spectra with broad (SWATH) or no precursor filtering. Retention time information of both MS1 precursors and MS2 fragments is then used to deconvolve the complex MS2 spectra and link each back to its precursors. In most cases, this requires DDA-derived MS2 spectrum libraries; however, there are now methods which allow full deconvolution and precursor identification without requiring a library!


So, I hear you ask, what are we at DC Biosciences doing about this? Well, for the time being, we are offering de novo sequencing for projects that require it – with the aim to offer it more systematically in future as the technology becomes more easily available. With our new Protein-from-RNA product, we can convert your RNA sequences into a bespoke proteome for your sample analysis. Lastly, keep an eye on our website, we are aiming to very soon start offering DIA analysis!

Point of View
Related Posts: