Does the colony-scoring marker change the reported reprogramming efficiency?

Answer

Only one verified source directly measured both stains on the same colonies, and it found no disagreement: all 23 hand-picked, morphologically stem-cell-like colonies from polycistronic-retroviral-OSKM-reprogrammed human fibroblasts were AP-positive and TRA-1-60-positive, a fold ratio of 1.0 50272361. That result cannot be generalized into "AP and TRA-1-60 always agree," because the colonies were pre-selected by eye for hESC-like morphology before either stain was applied — the comparison never included AP-positive colonies that looked partially reprogrammed, which is exactly the population where AP's known permissiveness would show up. Two further searches of the reachable literature, including a systematic six-candidate check, found no other source reporting matched AP+ and TRA-1-60+ colony counts on the same reprogramming material d4c5e9ff fbbf54e1.

Summary

The research question — does the scoring marker systematically change reported efficiency — cannot be answered from this corpus, and the reason is itself informative: papers that use both stains almost never print both counts side by side. The one verified head-to-head comparison agrees perfectly (1.0-fold), but only because it was run on pre-screened, already-morphologically-clean colonies, which is the one condition under which AP and TRA-1-60 would be expected to agree regardless of AP's general permissiveness. Readers should not treat the 1.0-fold figure as evidence that marker choice does not matter to published efficiencies in general.

Evidence

• 2011 PLoS ONE (retroviral OSKM + small-molecule cocktail, human fibroblasts) — sequentially scored the same 23 hESC-morphology-selected colonies for AP and TRA-1-60; all 23 were positive for both, fold ratio 1.0 (reported as "data not shown" in the original text). Colonies were picked by morphology before staining, so the comparison excludes AP-positive colonies that would not have passed visual pre-selection. 50272361
• 2018 blood-reprogramming paper — used both stains to pick an optimal scoring day, but only ever counted TRA-1-60+ colonies numerically; the AP result is a qualitative photograph, not a count, so no ratio is computable from it. d4c5e9ff
• Systematic check of six literature candidates — Chan et al. 2009 states AP is qualitatively "insufficient" versus TRA-1-60/DNMT3B/REX1 but full text (paywalled) could not confirm matched counts; Onder et al. 2012 reports AP+ only for mouse wells and TRA-1-60+ only for separate human wells (different material); Wang et al. 2019 reports a different ratio (homogeneous-vs-total TRA-1-60+ colonies); Park et al. 2012 reports TRA-1-81+/NANOG (a different marker than TRA-1-60) at 50-80% of AP+ colonies; Bharathan et al. 2017 reports TRA-1-60+ efficiency (~0.1%) with no AP+ count in the same study. None yields a computable AP-vs-TRA-1-60 fold ratio. fbbf54e1

Methodology

I treated "source reports both markers on the same material" as the entry condition for any usable data point, per the task's own required-measurement field (source_reports_both). Only one of the three verified submissions clears that bar. I did not treat the Park et al. or Wang et al. candidates surfaced by the literature-search fact as usable substitutes, because they compare a different marker pair (TRA-1-81/NANOG-vs-AP, or TRA-1-60-homogeneous-vs-TRA-1-60-total) than the AP-vs-TRA-1-60 pair this task asks about; naming them as near-misses is honest, counting them as data is not. I did not compute or estimate a fold ratio beyond the one already printed in the corpus.

Limitations

• The one usable data point (1.0-fold) was measured on colonies already pre-selected by eye for hESC-like morphology, so it cannot detect whether AP marks a larger population of partially reprogrammed colonies that a morphology screen, or a TRA-1-60 stain, would have excluded before counting even began. The paper's own colonies were also explicitly a "data not shown" aside, not the study's primary result.
• Twenty-three colonies from one study is too small and too specific (one delivery method, one small-molecule protocol, one cell type) to generalize a 1.0-fold ratio to retroviral, lentiviral, Sendai, or episomal reprogramming broadly.
• The corpus's second source demonstrates a structural problem beyond this one task: even when both stains are applied by the same lab on the same dishes, only one is typically reported as a number, which means the scoring-artefact question this task asks about is very likely under-documented across the wider published literature, not just under-verified in this corpus.