Phosphoproteomics: IMAC or TiO₂, and when FAIMS helps
When a phosphoproteomics run disappoints, the instrument is rarely the cause. Three decisions made before the sample reaches the LC determine most of the result.
Decision one: input
Phosphopeptides are a small fraction of a tryptic digest. Enrichment cannot create material that is not there, so deep phosphoproteomes start from more protein than global proteomes, typically hundreds of micrograms to a milligram per sample. The commonest reason a phospho experiment under-delivers is that it was planned with global-proteome input amounts.
Decision two: enrichment chemistry
| Chemistry | Strength | Watch for |
|---|---|---|
| Ti-IMAC / Zr-IMAC | High selectivity and capacity; good for mono-phosphopeptides at scale | Buffer composition matters; acidic loading with organic modifier |
| Fe-NTA IMAC | Widely available; good capacity | Some non-specific binding of acidic peptides |
| TiO₂ | Robust, inexpensive; tolerates variation | Competes for acidic peptides unless a competitor (glycolic/lactic acid) is used |
| Sequential IMAC → TiO₂ | Broader coverage of multiply and singly phosphorylated peptides | Twice the handling; more loss steps |
For groups setting up enrichment for the first time, the EasyPhos protocol (Humphrey et al., Nat Protoc 2018) is a well-documented single-step TiO₂ workflow in 96-well format that scales to hundreds of samples; it is a sensible benchmark against which any kit or in-house method can be compared.
Yield is not the right metric on its own. An enrichment with high yield but 40% unmodified peptides spends the gradient on the wrong molecules; selectivity, the fraction of identified peptides that are phosphorylated, is what predicts depth.
Decision three: separation before the analyser
Phosphopeptides co-elute with each other and with residual unmodified peptides, so co-isolation limits both identification and site localisation. Two remedies: off-line fractionation at basic pH, which costs instrument time, and gas-phase separation with FAIMS, which does not. FAIMS on hybrid Orbitrap instruments has been shown to increase phosphopeptide identifications in single-shot runs by reducing co-isolation (Muehlbauer et al., Anal Chem 2020). In our own method work the gain is consistent across compensation voltages chosen for phosphopeptides, and it is the first thing we add before extending a gradient.
Reporting the result
Most phosphosites detected in a deep experiment have no known function; a large fraction of the human phosphoproteome has been catalogued, but functional annotation lags far behind (Ochoa et al., Nat Biotechnol 2019). Prioritisation therefore matters as much as detection.
A site is only useful if it is localised. Each phosphosite should carry a localisation probability, and downstream analysis should be restricted to confidently localised sites. Running a global proteome from the same samples lets a site change be interpreted against its protein's abundance, so a site that merely tracks protein level is not mistaken for regulation. Kinase-substrate enrichment then turns the site list into a list of likely upstream kinases.
We run the complete workflow as a service and supply TiO₂ and magnetic-bead enrichment kits for groups who enrich in-house.
References
- Humphrey et al. High-throughput and high-sensitivity phosphoproteomics with the EasyPhos platform. Nat Protoc 2018. PubMed · DOI
- Muehlbauer LK et al. Global phosphoproteome analysis using high-field asymmetric waveform ion mobility spectrometry on a hybrid Orbitrap mass spectrometer. Anal Chem 2020;92:15959-15967. Open access PubMed · DOI · Full text (PMC)
- Ochoa et al. The functional landscape of the human phosphoproteome. Nat Biotechnol 2019. Open access PubMed · DOI · Full text (PMC)
Citations verified against PubMed records; “Open access” marks papers whose full text is freely available via PubMed Central or the journal.
Written by the omicslink proteomics team. Reviewed November 2026.
Discuss your phospho project