CharmeRT - Chimeric spectra identification utilizing retention time prediction - Up to 50% more peptide IDs

Co-elution of peptides is a major issue in bottom-up proteomics experiments, since it can led to co-isolation and co-fragmentation of peptide ions resulting in a chimeric MS2 spectra.  A MS2 spectrum containing more than a one peptide spectrum match (PSM) is considered to be chimeric MS2 spectra.

By the way chimera is taken as an analogy from greek mythology, where it stands for a hybrid creature from two or more animals. An example would be pegasus, the winged horse, so basically a mixture of a horse and a bird. 

But now back to the problem having multiple peptides selected as MS2 precursors… According to the study a HeLa digest separated by a 3h gradient displayed over 50% chimeric MS2 spectra, 20% of these spectra showed even more than two PSMs in a single MS2.

Sounds awesome to me and really is a reason to further investigate - how the authors have done this.

Well, as it turns out they identified the chimeric spectra by performing a 2^nd search, in which previously assigned PSMs have been removed from the raw data. These multiple PSMs were then validated utilizing retention time prediction model as part of Elutator software. This prediction model is based on the hydrophobicity index of the (modified) peptides. The index determines how much energy is needed to transfer molecules (or a peptide chain) from a nonpolar solvent to water.

If the polypeptide requires energy to do this it is hydrophobic. If it does not it is considered to be hydrophilic. In reverse phase liquid chromatography such an index can provide a measure of approximate elution order and time.

The special feature of Elutator is that it not only takes the hydrophobicity of the individual amino acids into account but also it considers the effect of neighbouring amino acids during its calculations. This lets Elutator predict retention times much more precise compared to other RT prediction tools.

CharmeRT using Elutator outperformed conventional proteomics workflows such as mascot and percolator by up 90% in terms of assigned PSMs and up to 65% for validated peptide IDs. It is claimed that most of these newly identified peptides are low abundant peptides.  Therefore the authors tried to investigate this hypothesis with RNA data of the identified peptides. And indeed, they showed an extension in sensitivity towards smaller peptide abundancies without increasing instrument time, well just post processing time, though.

Charme RT can also be beneficial for DIA approaches, since it works much better for wider quadrupole isolation windows.

Source: https://pubs.acs.org/doi/10.1021/acs.jproteome.7b00836