Correction to: Nature https://doi.org/10.1038/s41586-026-10182-7 Published online 4 March 2026
In the version of this article initially published, in the Fig. 4d Source Data and associated figure plot, due to an error during the full width at half maximum resolution analysis using the ‘fwhm_on_spots’ jython-fiji macro, one channel from one image was inadvertently loaded twice into the analysis script. The corrected analysis changes the PIEZO2 FWHM from 83 ± 31 nm to 88 ± 32 nm and the number of PIEZO2 puncta from n = 201 to n = 320. The figure panel and Source Data are now updated. In the Fig. 4d legend, the reported n value for FLNB puncta was incorrect and is now updated to n = 468 rather than 469.
In the Source Data for Figs. 1–3, the “Raw MINFLUX output” tabs for cell #2 of the “PIEZO2 TCOK105 – Isotonic” condition contained values from an analysis version that used a different standard-deviation-per-trace threshold from that used for the figures. Separately, in the “Raw MINFLUX output” tab in the Source Data of Fig. 1 and Extended Data Fig. 4 for cell #3 of the “mPIEZO1 TCO*K103 – Hypotonic” condition, an analysis output was uploaded in which two molecules that should have been excluded by the z-axis filtering criterion were inadvertently retained. (The Source Data Extended Data Fig. 4 file was originally mislabeled Source Data Extended Data Fig. 3).
Following the Source Data updates, the Fig. 1b legend now reports n = 102 instead of n = 52 molecules for the unstimulated PIEZO2 condition; Fig. 1d now reports a median of 37.1 nm and n = 39 molecules instead of 34.7 nm and n = 41 molecules; Fig. 1e now reports 20.8 nm instead of 19.9 nm for the unstimulated PIEZO2 condition; the PIEZO2 values in Fig. 3e now report median = 20.8 nm and n = 102 molecules rather than median = 18.0 nm and n = 30 molecules. In the Fig. 1–3 legends, in three cases, mean values were reported rather than the correct median values, where for Fig. 1h, PIEZO1 hyperosmotic condition, 19.8 nm has been updated to 17.1 nm; for Fig. 2b, PIEZO2 + cytochalasin D at rest, 20.1 nm has been updated to 18.0 nm; for Fig. 3e, Flnb DsiRNA + hypo-osmotic swelling, 37.0 nm has been updated to 38.7 nm. In the Fig. 1 and Extended Data Fig. 2 legends, sample sizes were listed incorrectly, where for Fig. 1k, PtK2 control cells, n = 6 has been updated to n = 7 cells; for Extended Data Fig. 2d, direct Alexa 647, n = 2,014 has been updated to 2,013 traces, fluorogenic DNA PAINT, n = 7,254 has been updated to 7,253 traces, and for traditional DNA PAINT, n = 1,013 has been updated to 1,012 traces. In the Extended Data Fig. 8c legend, a typographical error in “Piezo2 Flnb KO” has been corrected to “Piezo1 Flnb KO”.
The corrections do not affect the outcome or significance of any reported statistical comparison, the direction of any experimental result, or the scientific conclusions or interpretations presented in the article. Figure 4d, the legends for Figs. 1–4 and Extended Data Figs. 2 and 8, and Source Data for Figs. 1–4 and Extended Data Fig. 4 are now updated in the HTML and PDF versions of the article. For comparison, the original Fig. 4 is available as Supplementary Information accompanying this amendment.
Supplementary information is available in the online version of this amendment.
Author information
Authors and Affiliations
Corresponding authors
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/
About this article
Cite this article
Mulhall, E.M., Yarishkin, O., Hill, R.Z. et al. Author Correction: The molecular basis of force selectivity by PIEZO2. Nature (2026). https://doi.org/10.1038/s41586-026-11079-1
Published:
Version of record:
DOI: https://doi.org/10.1038/s41586-026-11079-1
Facts Only
* E.M. Mulhall, O. Yarishkin, R.Z. Hill, and colleagues authored the correction.
* The correction applies to a Nature article published online March 4, 2026.
* A jython-fiji macro error in Figure 4d resulted in one image channel being loaded twice.
* PIEZO2 FWHM in Figure 4d changed from 83 ± 31 nm to 88 ± 32 nm.
* PIEZO2 puncta count in Figure 4d changed from n = 201 to n = 320.
* FLNB puncta count in Figure 4d legend changed from n = 469 to n = 468.
* Source Data for Figures 1–3 (cell #2, PIEZO2 TCOK105 – Isotonic) used an incorrect standard-deviation-per-trace threshold.
* Source Data for Figure 1 and Extended Data Figure 4 (cell #3, mPIEZO1 TCO*K103 – Hypotonic) included two molecules that failed z-axis filtering.
* Figure 1b unstimulated PIEZO2 count changed from n = 52 to n = 102 molecules.
* Figure 1d reported values changed from 34.7 nm (n = 41) to 37.1 nm (n = 39).
* Figure 1e unstimulated PIEZO2 value changed from 19.9 nm to 20.8 nm.
* Figure 3e PIEZO2 values changed from median = 18.0 nm (n = 30) to median = 20.8 nm (n = 102).
* Median values replaced mean values in Figure 1h (19.8 nm to 17.1 nm), Figure 2b (20.1 nm to 18.0 nm), and Figure 3e (37.0 nm to 38.7 nm).
* Sample sizes were updated for Figure 1k (n = 6 to n = 7) and Extended Data Figure 2d.
* A typo in Extended Data Figure 8c was corrected from "Piezo2 Flnb KO" to "Piezo1 Flnb KO".
Executive Summary
A formal correction has been issued for a study on the molecular basis of force selectivity by PIEZO2, published in Nature. The amendments address technical errors in data processing, including a macro error that caused a single image channel to be loaded twice during full width at half maximum (FWHM) analysis, as well as inconsistencies in the standard-deviation-per-trace thresholds and z-axis filtering criteria used in the Source Data.
These errors necessitated updates to several figures and legends, specifically affecting reported sample sizes (n values), median measurements, and the identification of specific knockout conditions. While the numerical values for PIEZO2 FWHM and various molecule counts have shifted, the authors maintain that these changes do not alter the statistical significance of the comparisons, the direction of the experimental results, or the overall scientific conclusions and interpretations of the research.
Full Take
The correction reveals a series of systemic "bookkeeping" errors typical of high-resolution imaging pipelines where automated macros and multi-stage filtering are used. The most significant technical lapse—loading a single channel twice—highlights a vulnerability in the data-loading phase of the analysis script. While the authors assert that the scientific conclusions remain intact, a peer reviewer would note that the PIEZO2 puncta count in Figure 4d increased by over 50% (from 201 to 320), and the sample size in Figure 1b nearly doubled.
In scholarly work, such shifts in 'n' values can occasionally impact the power of statistical tests, though the reported change in FWHM (83 to 88 nm) is relatively minor. The transition from reporting mean values to median values in several figures suggests a correction toward a more appropriate measure of central tendency for potentially skewed biological data, which is a standard requirement for rigorous reporting.
The primary concern for a researcher would be the reliability of the "Raw MINFLUX output" tabs. The admission that different thresholds were used for the figures versus the source data indicates a disconnect between the analysis pipeline and the archival record.
To strengthen the validity of these findings, a third-party replication of the FWHM analysis using the corrected raw data would be the logical next step.
Bridge Questions:
1. Does the increase in sample size (n) for specific conditions fundamentally change the p-values of the primary comparisons?
2. If the analysis script had a loading error in one instance, what validation steps were implemented to ensure other scripts were not similarly flawed?
3. How sensitive are the overall conclusions to the shift from mean to median values in the reported data?
Counterstrike Scan: This is a standard academic correction of technical errors. The content does not match any known influence campaign patterns; it is a transparent, though granular, admission of data-handling mistakes.