Skip to content

Relevant vs. confounding proteins

ProteoformTracker runs two genuinely different searches, matching a distinction from its design spec:

  • Relevant proteoforms (Section 1): the isoforms/proteoforms you explicitly checked — proteins that share exon structure or sequence with your target, by construction (you picked them from the same gene, or PTM-derived variants of the same sequence).
  • Confounding proteins (Section 2): proteins that share mass and/or m/z with your target regardless of any sequence or gene relationship — found by searching the full reference human proteome, not by anything you selected.

Cross-gene proteins essentially never share extended sequence identity, so their fragment ions diverge quickly at MS2 even when their intact mass happens to coincide at MS1 — which is exactly why the confounder search is a real risk worth checking explicitly, not something you'd catch just by comparing your relevant proteoforms against each other.

Two ways a protein can confound your target

The confounder search unions two independent searches, since they catch genuinely different risks:

  • Mass-domain: the candidate's own intact mass falls within your target's resolving-power window (see Resolving power & FWHM) — a candidate found this way could still turn out resolvable once real charge-state peaks are checked.
  • m/z-domain: the candidate's charge-state peaks land on top of your target's, even when its intact mass is far away — e.g. a 12 kDa and a 16 kDa protein can produce colliding peaks at a 4:3 charge-state ratio. A mass-window-only search would never surface this.

Each candidate in the results table is tagged found_via: mass, mz, or both.

Why the list is capped

An m/z-domain search alone can surface hundreds of real hits for some targets — every reference protein whose charge envelope happens to touch the target's anywhere across its whole envelope. ProteoformTracker caps the shown list (default: the 30 highest-priority), prioritized both > mass-only > mz-only-sorted-by-collision-severity — never an arbitrary or first-found subset. The status text reports how many were found in total vs. how many are shown.

The MS2 shared-ions column

Before you decide which confounders are even worth the (isotope-computation-heavy) full comparison, the candidate table shows a cheap pairwise MS2 overlap count per candidate: how many of the target's own b/y ions collide with that one candidate's fragment ladder, considered alone. This is deliberately not the same thing as the unique/partial/common tiers shown after comparing — those are inherently set-dependent (a fragment only counts as "common" if every confounder you leave checked shares it), so they can't be known until you've actually picked your final comparison set. The pairwise number is there purely to help you prioritize which candidates are worth including.

Two-step workflow, by design

Section 2 is deliberately split into a cheap search step (candidate list, no isotope computation) and an expensive compare step (only for whichever candidates you leave checked) — see Section 2: Confounding-protein search for the full workflow, including why this split exists and how to use outside evidence (e.g. RNA-seq expression calls) to narrow the list before running the heavy computation.