MHz compression of pulse packets facilitates remote focusing of electroporation (DANDI:001047)

Claims

From the paper MHz compression of pulse packets facilitates remote focusing of electroporation

Reproducing Silkuniene et al. (2025) from dandiset 001047 9

Notebook reproduce_silkuniene_2025.ipynb

Partially reproduced1

Partially reproducedResults 3.1, paragraph 4; Figure 4Bverifier disagreed at firstrests on an assumption
YP uptake at the array centre showed no significant effect of packet rate.
YP intensity in the center showed no significant effect of the packet rate ( t -test with Bonferroni correction, Fig. 4B ).
Results 3.1, paragraph 4; Figure 4B
At x1 (2 bursts, adopted) centre YP does not differ between rates, 1 Hz 1.0 +/- 21.8 against 0.2 MHz -1.4 +/- 6.3 a.u., t = 0.09, p = 0.93 (n = 6 and 4), uncorrected, so no Bonferroni factor (1, 4 or 10) changes it. The paper's test finds no effect, but neither rate stains the centre at all, where the paper's centre is stained at every rate, and only one of the paper's comparisons (no 1 kHz or 10 kHz) can be made; the outcome changed from reproduced after verification for these reasons. Under the total reading 1 Hz x2 (2 bursts, 516 a.u.) is above 0.2 MHz x1 (1 burst, p = 0.007, 0.03 with a factor of 4) and not different from 0.2 MHz x3 (3 bursts, 353 a.u., p = 0.64). In the x5 variant 1576 +/- 465 against 2275 +/- 922 a.u., p = 0.54. Without the flagged 1 Hz area 013Control-x1_obj-1pihpj7 R, 4.4 +/- 26.4 against -1.4 +/- 6.3 a.u., p = 0.85 (n = 5 and 4).
How the test differed from the paper’s method
- Two rates, so one comparison under both readings of the claims file. Burst labels read as repeats after the first burst, so that x1 is the paper's 2 bursts (x4) at both rates (assumption). The files' protocol text says the protocol "was either repeated 2, 3, or 5 times" upon completion, and under this reading x5 is 6 bursts (x4), the paper's largest number (Figure 7). Under the other reading (the label is the total) 2 bursts exists only at 1 Hz (x2); that reading is shown, with 1 Hz x2 against 0.2 MHz x1 and x3, which bracket 2 bursts. x5 is shown as a variant and decides nothing.

Not reproduced8

Not reproducedResults 3.1, paragraph 2; Figure 3Arests on an assumption
Every CANCAN protocol enhanced electroporation at the array centre, while in the control (single unipolar pulse per packet) YP uptake decreased gradually toward the centre.
All CANCAN protocols visibly enhanced electroporation at the center of the array, in contrast to its gradual decrease toward the center in the control ( Fig. 3A ).
Results 3.1, paragraph 2; Figure 3A
At x1 (2 bursts, adopted) no CANCAN profile peaks at the centre under any reading. The 1 Hz profile peaks 6.5 to 7 mm from it (n = 6) and the 0.2 MHz wells are essentially unstained (centre -1.4 a.u., maximum 21 to 37 a.u., n = 4). Under the total reading the 1 Hz x2 profile (2 bursts, n = 2) peaks 5 to 6 mm from the centre, so the CANCAN part fails under both readings of the labels. The pooled x1 control is lower at the centre (245 +/- 172 a.u.) than at 6 mm (868 +/- 273) and 7 mm (618 +/- 197), O1 a, O2 a, without a gap beyond the error bars under any F8 reading. In the x5 variant no CANCAN profile peaks at 0. Without the flagged 1 Hz control area 013Control-x1_obj-6stwj8 W, whose centre sits on the stain, the control part holds under every reading (centre 74 +/- 23 a.u. against 990 +/- 268 at 6 mm and 668 +/- 209 at 7 mm, n = 11, O1 a, O2 a), but the CANCAN part fails as before, so the outcome is the same.
How the test differed from the paper’s method
- Only 1 Hz and 0.2 MHz are in the dandiset; the paper's other rates (1 kHz, 10 kHz and one more below 10 kHz) are not.
- Burst labels read as repeats after the first burst, so that x1 is the paper's 2 bursts (x4) at both rates (assumption). The files' protocol text says the protocol "was either repeated 2, 3, or 5 times" upon completion, and under this reading x5 is 6 bursts (x4), the paper's largest number (Figure 7). Under the other reading (the label is the total) 2 bursts exists only at 1 Hz (x2); that reading is shown, with 1 Hz x2 against 0.2 MHz x1 and x3, which bracket 2 bursts. x5 is shown as a variant and decides nothing. The paper's controls are of unstated rate; the controls of both rates at the same label are pooled, and each rate is shown alone as a variant.
- The error-bar rule (F8, Methods 2.5, "A gap exceeding the length of the error bars") is read as a gap beyond the longer s.e. arm, so that the gap exceeds each error bar (assumption); the mean-arm and shorter-arm readings are shown.
- The array centre is found by matching the paper's electrode geometry (Methods 2.2) to the footprints in the YP image, or, where the match is weak, from the centroid of the stain (an assumption of four-fold symmetry); the paper placed ROIs by hand. Six areas whose centre is visibly misplaced are flagged and kept, and the claims they enter are also shown without them, with the same outcomes. Three areas whose centre fell near an image edge are dropped.
- Background is the mean over the four footprints of 3 ROIs each (Methods 2.5 says 3 ROIs in one footprint without saying which); the darkest footprint is shown as a variant for the ratio claim.
Not reproducedResults 3.1, paragraph 2; Figure 3B, 3Crests on an assumption
Raising the packet rate from 1 Hz to 0.2 MHz suppressed electroporation better both near the electrodes (diagonal profile) and at the perimeter between electrodes (profile parallel to the sides).
The most conspicuous effect of increasing the packet rate from 1 Hz to 0.208333 MHz (abbreviated to 0.2 MHz below and in figure legends) was better suppression of electroporation both near electrodes ( Fig. 3B ) and at the perimeter between electrodes ( Fig. 3C ).
Results 3.1, paragraph 2; Figure 3B, 3C
At x1 (2 bursts at both rates, adopted) YP at 0.2 MHz is below 1 Hz at 6 mm (9.1 +/- 3.8 against 152.5 +/- 76.2 a.u., t test p = 0.16 to 0.17) and at 7 mm (4.9 +/- 2.9 against 174.0 +/- 130.9, p = 0.33 to 0.34), with a gap beyond the longer error bar at neither; the mean of the parallel ROIs at 4 mm or more is higher at 0.2 MHz (6.8 against -10.1 a.u., O1 a), so the claim fails at x1 under every F8 reading. Under the total reading there is no 0.2 MHz 2-burst group. 1 Hz x2 (2 bursts) is above 0.2 MHz x3 (3 bursts) at 7 mm and at the perimeter with a gap beyond the longer error bar, and at 6 mm only under the shorter-arm reading (731 against 88 a.u., p = 0.28 to 0.29, n = 2 and 2), so the brackets point the paper's way with two wells per group. In the x5 variant 0.2 MHz is higher than 1 Hz at 6 and 7 mm (1534 against 1173, 917 against 780). Without the flagged 1 Hz area 013Control-x1_obj-1pihpj7 R (n = 5 and 4), the x1 gap exceeds the longer arm at 6 mm (183.6 against 9.1 a.u., p = 0.11) and not at 7 mm (p = 0.28), and perimeter (b) is still higher at 0.2 MHz, so the claim fails at x1 under every F8 reading.
How the test differed from the paper’s method
- Burst labels read as repeats after the first burst, so that x1 is the paper's 2 bursts (x4) at both rates (assumption). The files' protocol text says the protocol "was either repeated 2, 3, or 5 times" upon completion, and under this reading x5 is 6 bursts (x4), the paper's largest number (Figure 7). Under the other reading (the label is the total) 2 bursts exists only at 1 Hz (x2); that reading is shown, with 1 Hz x2 against 0.2 MHz x1 and x3, which bracket 2 bursts. x5 is shown as a variant and decides nothing.
- The error-bar rule (F8, Methods 2.5, "A gap exceeding the length of the error bars") is read as a gap beyond the longer s.e. arm, so that the gap exceeds each error bar (assumption); the mean-arm and shorter-arm readings are shown.
- The parallel profile is measured on every area; the paper measured it in one set of experiments, which the dandiset does not identify.
Not reproducedResults 3.1, paragraph 3; Figure 3B, 3C legend; Methods 2.5rests on an assumption
The decline of YP emission from the centre was bell-shaped and fitted a Lorentzian function over -4 to 4 mm nearly perfectly (R > 0.9, p < 0.001) for every packet rate.
The fit correlation with data in the −4 to 4 mm radius was nearly perfect (R>0.9, p<0.001).
Results 3.1, paragraph 3; Figure 3B, 3C legend; Methods 2.5
What the paper reports
- correlation of the Lorentzian fit with the data, -4 to 4 mm: R > 0.9, p < 0.001
- significance of each YP fit in Figures 3, 5 and 6: p < 10⁻⁴
Measured is R for 1 Hz CANCAN at x1, diagonal, O1 a, O2 a, the first reading listed. At x1 (2 bursts, adopted) the 1 Hz fits give R from 0.36 to 0.72, and the 0.2 MHz profiles have no central peak (all 8 fits with the width at the 100-mm bound, R from -0.74 to -0.02). Under the total reading the 1 Hz x2 fits give R from 0.40 to 0.94 and pass in 2 of 8 profile and reading combinations. No fit failed; 12 of 48 ran to the width bound. 0.2 MHz x3 and x5 pass in 4 of 8 combinations each (R 0.79 to 1.00 and 0.80 to 0.96); 1 Hz x5 gives R from -0.89 to 0.51. Without the flagged 1 Hz area 013Control-x1_obj-1pihpj7 R, the x1 1 Hz fits give R from 0.39 to 0.74, none passing.
How the test differed from the paper’s method
- Only 1 Hz and 0.2 MHz. Burst labels read as repeats after the first burst, so that x1 is the paper's 2 bursts (x4) at both rates (assumption). The files' protocol text says the protocol "was either repeated 2, 3, or 5 times" upon completion, and under this reading x5 is 6 bursts (x4), the paper's largest number (Figure 7). Under the other reading (the label is the total) 2 bursts exists only at 1 Hz (x2); that reading is shown, with 1 Hz x2 against 0.2 MHz x1 and x3, which bracket 2 bursts. x5 is shown as a variant and decides nothing.
Not reproducedResults 3.1, paragraph 3; Figure 3B, 3Cverifier disagreed at firstrests on an assumption
Beyond the fitted range, the data deviated from the Lorentzian fit less as the packet rate increased.
The deviation of experimental data from the fits at larger distances from the center diminished with increasing the packet repetition rate, highlighting better suppression of the off-target electroporation.
Results 3.1, paragraph 3; Figure 3B, 3C
At x1 (2 bursts, adopted) the mean deviation beyond 4 mm is larger at 1 Hz than at 0.2 MHz under every reading (diagonal, O1 a, 145.0 against -3.7 a.u.), but every 0.2 MHz fit has its width at the 100-mm bound (fig3-lorentzian-profile), so there is no fitted peak for the data to deviate from; the order shows only that the 0.2 MHz wells are unstained far out, as everywhere. The outcome changed from partially reproduced after verification, which pointed this out. Under the total reading there is no 0.2 MHz 2-burst group; 1 Hz x2 against 0.2 MHz x3 (3 bursts), both fits with a peak, gives the paper's order under every reading (diagonal, O1 a, 486.8 against 75.6 a.u.). In the x5 variant the order is reversed under every reading (-207.5 against 429.9 a.u.). With two rates the Spearman coefficient is +1 or -1 and has no test.
How the test differed from the paper’s method
- Only 1 Hz and 0.2 MHz, so a trend over rates reduces to one comparison. Burst labels read as repeats after the first burst, so that x1 is the paper's 2 bursts (x4) at both rates (assumption). The files' protocol text says the protocol "was either repeated 2, 3, or 5 times" upon completion, and under this reading x5 is 6 bursts (x4), the paper's largest number (Figure 7). Under the other reading (the label is the total) 2 bursts exists only at 1 Hz (x2); that reading is shown, with 1 Hz x2 against 0.2 MHz x1 and x3, which bracket 2 bursts. x5 is shown as a variant and decides nothing.
- A fit whose width runs to its bound is taken as having no peak, so that no deviation from a bell-shaped fit can be measured; the order of the deviations alone is also shown.
Not reproducedResults 3.1, paragraph 4; Figure 4Brests on an assumption
In contrast to the centre, YP uptake at 6 and 7 mm from the centre (1.65 and 0.65 mm from the electrodes) declined steeply as the packet rate increased.
It contrasted with a steep, packet rate-dependent decline of YP uptake at 6 and 7 mm from the center (i.e., 1.65 and 0.65 mm from electrodes).
Results 3.1, paragraph 4; Figure 4B
At x1 (2 bursts, adopted) the power exponent is negative at 6 and 7 mm (k = -0.23 and -0.29, the same for means and per-area values) and the 0.2 MHz means are lower (9.1 against 152.5 and 4.9 against 174.0 a.u.), but the gap exceeds the longer error bar at neither distance (t test p = 0.17 and 0.33). The outcome rests on the F8 reading: under the mean-arm reading the gap holds at 6 mm only, and under the shorter-arm reading at both, which would make the claim reproduced. On two-well images only (leaving out the whole-well images, on another a.u. scale) the gap exceeds the longer arm at 6 mm (79.8 against 9.1 a.u., p = 0.06) and not at 7 mm. Under the total reading, 1 Hz x2 against 0.2 MHz x3 has k < 0 at both distances, with a gap beyond the longer arm at 7 mm only. In the x5 variant the exponent is positive at both distances (0.022 and 0.013). Without the flagged 1 Hz area 013Control-x1_obj-1pihpj7 R the result is the same: k = -0.25 and -0.31, a gap beyond the longer arm at 6 mm (183.6 against 9.1 a.u., p = 0.11) and not at 7 mm (210.4 against 4.9, p = 0.28), and the claim holds only under the shorter-arm reading.
How the test differed from the paper’s method
- With two rates the power function passes through both means. Burst labels read as repeats after the first burst, so that x1 is the paper's 2 bursts (x4) at both rates (assumption). The files' protocol text says the protocol "was either repeated 2, 3, or 5 times" upon completion, and under this reading x5 is 6 bursts (x4), the paper's largest number (Figure 7). Under the other reading (the label is the total) 2 bursts exists only at 1 Hz (x2); that reading is shown, with 1 Hz x2 against 0.2 MHz x1 and x3, which bracket 2 bursts. x5 is shown as a variant and decides nothing.
- The error-bar rule (F8, Methods 2.5, "A gap exceeding the length of the error bars") is read as a gap beyond the longer s.e. arm, so that the gap exceeds each error bar (assumption); the mean-arm and shorter-arm readings are shown.
Not reproducedResults 3.1, the sentences on the ratio of YP uptake in the center and near electrodes; Figure 4B; Abstractrests on an assumption
At 0.2 MHz, YP uptake at the centre was 40 times that near the electrodes, despite the field there being about 4 times weaker (3.3 against about 14 kV/cm).
Now, by increasing the packet repetition rate to 0.2 MHz, we achieved a 40-fold difference ( Fig. 4B ), despite using a bigger array with larger reduction of the electric field from electrodes to the center.
Results 3.1, the sentences on the ratio of YP uptake in the center and near electrodes; Figure 4B; Abstract
Measured is the x1 ratio at 6 mm, O1 a, O2 a, the first reading listed. No label and reading gives 36 to 44. At x1 (2 bursts, adopted) the centre is below background (-1.4 a.u.) and the ratio is -0.29 to -0.15; with the darkest footprint as background, 0.53 and 0.65. Under the total reading, where 2 bursts lies between x1 and x3, x3 gives 2.55 to 3.99. In the x5 variant 1.44 to 2.48 (centre 2275 a.u., 6 mm 1534, 7 mm 917).
How the test differed from the paper’s method
- Burst labels read as repeats after the first burst, so that x1 is the paper's 2 bursts (x4) at both rates (assumption). The files' protocol text says the protocol "was either repeated 2, 3, or 5 times" upon completion, and under this reading x5 is 6 bursts (x4), the paper's largest number (Figure 7). Under the other reading (the label is the total) 2 bursts exists only at 1 Hz (x2); that reading is shown, with 1 Hz x2 against 0.2 MHz x1 and x3, which bracket 2 bursts. x5 is shown as a variant and decides nothing.
- Background is the mean over the four footprints of 3 ROIs each (Methods 2.5 says 3 ROIs in one footprint without saying which); the darkest footprint is shown as a variant for the ratio claim.
Not reproducedResults 3.2, paragraph 1; Figure 5A, 5Brests on an assumption
The central spot of YP staining (electroporation) was only slightly larger than the spot of Pr staining (cell killing).
Somewhat unexpectedly, the central spot of YP staining (electroporation) was just slightly larger than the spot of Pr staining (cell killing; Fig 5A and B ).
Results 3.2, paragraph 1; Figure 5A, 5B
At x1 (2 bursts, adopted) there is one usable 0.2 MHz CANCAN well with YP and Pr images; its YP profile has no spot (the Lorentzian width runs to the 100-mm bound under every reading), and its Pr area is flagged as misplaced, without which no x1 well is left. So at the S2 condition there is no YP spot to compare. Under the total reading there is no 0.2 MHz 2-burst well. At every other label the YP spot is narrower than the Pr spot under every reading, ratio x3 0.68 to 0.69, x5 0.52 to 0.66 (0.50 to 0.59 without the flagged Pr area), pooled 0.65 to 0.76 (0.644 to 0.725 without the flagged areas), all outside 1 to 1.5. The FWHM values (YP 4.8 to 8.3 mm, Pr 7.1 to 14.0 mm at x3 and x5) are far wider than the paper's (1.1 to 4 mm).
How the test differed from the paper’s method
- The dandiset gives no elevation. Assumed that the wells imaged for Pr had the electrodes touching the monolayer, the paper's default setting (Methods 2.2, "positioned ... until touching the bottom of the culture dish"). The data cannot check this.
- Burst labels read as repeats after the first burst, so that x1 is the paper's 2 bursts (x4) at both rates (assumption). The files' protocol text says the protocol "was either repeated 2, 3, or 5 times" upon completion, and under this reading x5 is 6 bursts (x4), the paper's largest number (Figure 7). Under the other reading (the label is the total) 2 bursts exists only at 1 Hz (x2); that reading is shown, with 1 Hz x2 against 0.2 MHz x1 and x3, which bracket 2 bursts. x5 is shown as a variant and decides nothing. The x3, x5 and pooled wells are shown; pooling mixes burst numbers whose centre YP differs over 1000-fold and is not the S2 condition.
- One x1 Pr area whose centre fell at the image edge is dropped. Pr areas are located in their own images; two of the five usable ones (one by the footprint match, one by the centroid) have visibly misplaced centres and are flagged, and results are shown without them.
Not reproducedResults 3.2, paragraph 2; Figure 5Brests on an assumption
Pr intensity declined from the peak as a Gaussian and YP as a Lorentzian; for the 0-mm data (the left panel of Figure 5B, read as the 0-mm panel) the R2 was 0.92 (Lorentzian) against 0.96 (Gaussian) for Pr, and 0.98 against 0.95 for YP, fitted from -5 to 5 mm with p < 0.0001.
For Pr data in the left panel of Fig. 5B , the coefficient of determination increased from 0.92 for Lorentzian fit to 0.96 for Gaussian fit. YP data showed the opposite trend, with the coefficients of 0.98 and 0.95 for Lorentzian and Gaussian fits, respectively (for all these fits, the fitted region was from −5 to 5 mm, p<0.0001).
Results 3.2, paragraph 2; Figure 5B
What the paper reports
- each of the four fits: p < 10⁻⁴
Measured are the x1 values under O1 a, O2 a (Pr Lorentzian, Pr Gaussian, YP Lorentzian, YP Gaussian). At x1 (2 bursts, adopted, one well) all four R2 are between -0.011 and 0.010 under every reading: neither profile has a shape either function fits. At x3 no reading has all four within tolerance and the Pr ordering fails under every reading. At x5 all four are within tolerance with both orderings under O1 a, O2 b only, and without the flagged Pr area the Pr ordering fails under every reading. Pooled, R2 is 0.77 to 0.93 (Pr, Lorentzian), 0.74 to 0.89 (Pr, Gaussian), 0.68 to 0.88 (YP, Lorentzian) and 0.62 to 0.80 (YP, Gaussian), no reading within tolerance, and the Pr ordering fails.
How the test differed from the paper’s method
- Same wells and assumptions as fig5ab-yp-spot-slightly-larger-than-pr.

Not tested in a notebook 11

9 data missing from the dataset, 2 metadata insufficient

Blocked11

BlockedMethods 2.3, paragraph 1; Figure 2· Data missing from the dataset
With packets at the maximum rate, heating followed the field and was largest near the electrodes; after 2 bursts it was confined to about 2 mm next to the electrodes and could exceed 30 C but not 44 C; after 6 bursts it spread further, exceeding 30 C even near the centre, yet never exceeded 44 C.
With 2 bursts ( Fig. 2A ), heating was restricted to an approximately 2-mm zone next to electrodes, where it could reach over 30 °C but did not exceed 44 °C. With 6 bursts ( Fig. 2B ), heating was more widespread ... Temperature could exceed 30 °C even near the center of the array, but did not exceed 44 °C even near the electrodes.
Methods 2.3, paragraph 1; Figure 2
The dandiset holds no thermochromic film images or movies; its 88 files are stitched fluorescence images of cell monolayers (DAPI, FITC and Cy3 channels).
BlockedResults 3.1, paragraphs 5 and 6; Figure 4C· Data missing from the dataset
Higher packet rates focused electroporation into a smaller spot; the FWHM fell from about 4 mm at 1 Hz to about 2 mm at 1 kHz, 10 kHz and 0.2 MHz (p = 0.03, 0.02 and 0.09 against 1 Hz), and could be under 15% of the inter-electrode gap.
FWHM decreased from ~4 mm at 1 Hz to ~2 mm at 1 kHz, 10 kHz, and 0.2 MHz packet rates (p = 0.03; 0.02, and 0.09, respectively; t -test with Bonferroni correction).
Results 3.1, paragraphs 5 and 6; Figure 4C
What the paper reports
- 1 kHz against 1 Hz: t-test, p = 0.03, Bonferroni-corrected (n = 3 experiments per rate)
- 10 kHz against 1 Hz: t-test, p = 0.02, Bonferroni-corrected (n = 3 experiments per rate)
- 0.2 MHz against 1 Hz: t-test, p = 0.09, Bonferroni-corrected (n = 3 experiments per rate)
The criterion compares 1 kHz and 10 kHz with 1 Hz; the dandiset has no images at 1 kHz or 10 kHz. The per-area Lorentzian FWHM at the two rates present is printed in the notebook for the record; many fits run to their bounds (6 of 10 at x1, 3 of 11 at x5).
BlockedResults 3.1, paragraph 5; Figure 4C legend· Data missing from the dataset
FWHM fell logarithmically with packet rate from 1 Hz to 10 kHz (R = 0.77, p < 0.001; n = 15), with no further narrowing from 10 kHz to 0.2 MHz.
FWHM decrease followed the logarithmic dependence from 1 Hz to 10 kHz (correlation coefficient R = 0.77, p<0.001), but there was no further focality increase when the packet rate increased from 10 kHz to 0.2 MHz.
Results 3.1, paragraph 5; Figure 4C legend
What the paper reports
- log fit of FWHM against packet rate, 1 Hz to 10 kHz: R = 0.77, p < 0.001 (n = 15 experiments (0.2 MHz excluded))
- slope of the log fit: 2-sided, one-sample t test, p < 10⁻⁴ (n = 15 experiments)
The log fit spans five rates from 1 Hz to 10 kHz; the dandiset holds only 1 Hz and 0.2 MHz.
BlockedResults 3.2, paragraph 1; Figure 5A, 5B· Metadata insufficient
Since with MHz-CANCAN there was practically no YP staining near electrodes (a finding the text cites to Figure 4), the only clear effect of raising the electrodes 1 or 2 mm above the monolayer was a weakening of electroporation in the central spot.
With MHz-CANCAN (CT26 cells, 2 bursts (x4), Fig. 4 ), there was practically no YP staining near electrodes to start with, so the only clear effect of raising electrodes was the weakening of electroporation in the central spot.
Results 3.2, paragraph 1; Figure 5A, 5B
No file or metadata field gives an electrode elevation, so wells exposed with raised electrodes, if any are in the dandiset, cannot be identified. The reason changed from data_missing after verification. The dandiset's sets (two rates, CANCAN and control, burst labels) do not show the design of set S2 (0.2 MHz at 0, 1 and 2 mm).
BlockedResults 3.2, paragraph 1; Figure 5C· Data missing from the dataset
In the central spot cells showed bright nuclear YP staining, indicating irreversible damage, while farther out reversibly electroporated cells were weakly stained with less contrast between nucleus and cytoplasm; cells inside the electrode footprint were unstained though present (Hoechst).
Indeed, a closer examination of the central spot confirmed bright YP staining of cell nuclei ( Fig. 5C , zone 1), which is indicative of the irreversible membrane damage and cell death. Reversible electroporation is distinguished by a much weaker YP staining and smaller difference in staining of the nucleus and cytoplasm, such as in zones 2 and 3 in Fig. 5C .
Results 3.2, paragraph 1; Figure 5C
The criterion needs full-resolution YP and Hoechst images of the S2 wells. The dandiset has 12 CT26 Hoechst images, one per plate pair with the YP and Pr images of May 24, but they were acquired at 10:35, before the YP (12:49) and Pr (14:24) images, consistent with the paper's Hoechst images before exposure; there is no Hoechst image after exposure. All these images are two-well images at an inferred 3.25 um per pixel, half the resolution of the 4x camera (1.625 um, stated in the CHO files), so a 2 to 4 um cytoplasm ring is about one pixel. No file gives the elevation. Verification read the block as metadata_insufficient; the reason is kept as data_missing because the criterion's own block applies (the full-resolution images are not in the dandiset).
BlockedResults 3.3, paragraph 2; Figure 6B· Data missing from the dataset
The central YP peak was highest for 900-ns pulses, but not significantly different from 300- or 600-ns pulses.
The central peak of YP uptake was higher for 900-ns pulses, but the difference from the other two pulse durations was not statistically significant (CT26 cells, 2 bursts (x4), Fig. 6B ).
Results 3.3, paragraph 2; Figure 6B
Every file describes 600-ns pulses; there are no 300- or 900-ns exposures.
BlockedResults 3.3, paragraph 2; Figure 6C· Data missing from the dataset
Normalized to the central peak, the sideways decline of YP was similar for the three pulse durations.
When the peak YP emission was taken as 100%, its decline sideways was similar for the three tested pulse durations ( Fig. 6C ).
Results 3.3, paragraph 2; Figure 6C
Every file describes 600-ns pulses; there are no 300- or 900-ns exposures.
BlockedResults 3.3, paragraph 2; Figure 6D legend· Data missing from the dataset
Longer pulses focused electroporation into a smaller spot; the FWHM fell with pulse duration (slope p = 0.003), 600- and 900-ns pulses differed from 300 ns, and 900-ns packets gave FWHM 1.7 +/- 0.08 mm, 11.1 +/- 0.6% of the gap between electrodes.
However, a more detailed analysis based on fitting of individual experiments (as described in 3.1 ) revealed significant sharpening of the central peak when packets were built from 600- and 900-ns pulses ( Fig. 6D ). The slope coefficient of the logarithmic fit of the data was different from zero at p = 0.003 (one sample, 2-sided t -test). 900-ns pulse packets reduced FWHM to 1.7 ± 0.08 mm, that was 11.1 ± 0.6% of the gap between electrodes.
Results 3.3, paragraph 2; Figure 6D legend
What the paper reports
- slope of FWHM against pulse duration: one sample, 2-sided t-test, p = 0.003 (n = 5 experiments per duration)
- one of 600 or 900 ns against 300 ns (the legend marks * and ** without the text naming which): 2-sided t-test, p = 0.028, Bonferroni-corrected (n = 5 experiments per duration)
- the other of 600 or 900 ns against 300 ns: 2-sided t-test, p = 0.0072, Bonferroni-corrected (n = 5 experiments per duration)
Every file describes 600-ns pulses; there are no 300- or 900-ns exposures.
BlockedResults 3.4, paragraph 2; Figure 7B, 7E legend· Data missing from the dataset
The central YP peak grew linearly with the number of bursts in both CT26 and CHO cells (slope p < 0.001), except that it saturated after 6 bursts (x4) in CT26 cells.
Increasing the number of bursts increased the amplitude of the central peak of YP uptake in both CT26 and CHO cells ( Fig 7A and D ). The increase was linear in both cell lines, except when it reached saturation after 6 bursts (x4) in CT26 cells ( Fig. 7B and E ).
Results 3.4, paragraph 2; Figure 7B, 7E legend
What the paper reports
- slope of central YP against bursts, CT26 (6 bursts (x4) excluded): 2-sided one-sample t test, p < 0.001 (n = 3 to 5 experiments per burst number)
- slope of central YP against bursts, CHO: 2-sided one-sample t test, p < 0.001 (n = 4 experiments per burst number)
The claim needs burst series in CT26 and CHO cells; all CHO files carry one burst label ("repeated 1 time"), so the CHO half is missing. The CT26 half was tested on the 0.2 MHz CANCAN wells and decides no outcome. Central YP rises with the label (-1.4, 353 and 2275 a.u. at x1, x3, x5), but the slope over the fitted points is not significant: 177 a.u. per burst over 2 and 4 bursts, p = 0.085 (6 wells), with the labels read as repeats after the first, and 583 a.u. per burst over 1, 3 and 5 bursts, p = 0.051 (12 wells), read as totals. Under the first reading the 6-burst mean (2275) is far above the fit extrapolated to 6 bursts (707 +/- 748), the reverse of saturation.
BlockedResults 3.4, paragraph 2; Figure 7C, 7F legend· Data missing from the dataset
The FWHM of the central peak increased in proportion to the number of bursts in CT26 cells (slope p < 0.001) but not in CHO cells (p = 0.44).
Concurrently, FWHM of the central peak increased proportionally to the burst number in CT26 cells, but not in CHO cells ( Fig. 7C and F ).
Results 3.4, paragraph 2; Figure 7C, 7F legend
What the paper reports
- slope of FWHM against bursts, CT26: 2-sided one-sample t test, p < 0.001
- slope of FWHM against bursts, CHO: 2-sided one-sample t test, p = 0.44
As for fig7be-peak-grows-with-bursts, there is no CHO burst series. The CT26 half was tested and decides no outcome: over both label readings, three fit limits, both O2 and O4 readings, with the 6-burst point in or out and fits at the width bound kept or dropped, no slope of FWHM against bursts is positive with p < 0.05 (0 of 72); most are negative. Mean FWHM is about 20 to 68 mm.
BlockedResults 3.4, paragraph 3; Figure 7A, 7C· Metadata insufficientverifier disagreed at first
One burst (x4) in CT26 cells gave the sharpest peak of the study, FWHM 1.1 +/- 0.1 mm, 7.3% of the inter-electrode gap and 2.3% of the array area, with YP near the electrodes 40 to 50 times smaller than at the centre.
That being said, 1 burst (x4) in CT26 cells produced the sharpest peak of electroporation seen in this study ... at FWHM of 1.1 ± 0.1 mm, it occupied just 7.3% of the inter-electrode gap and 2.3% of the electrode array area. Off-target effects, characterized by YP emission near electrodes, were 40–50 times smaller than at the remote target.
Results 3.4, paragraph 3; Figure 7A, 7C
The claim is CT26 only; it depends on fig7cf-fwhm-bursts for the FWHM method, not on its CHO half. With the labels read as repeats after the first burst (adopted for every claim) there is no 1-burst well (x1 is 2 bursts), so the claim is blocked for the burst labels; the reason changed from data_missing after verification. Read as totals, x1 would be 1 burst, and the result is far from the claim, so it would be not reproduced: mean FWHM 55 to 68 mm (2 of 4 per-well fits at the width bound), 358 to 445% of 15.3 mm, against 1.1 +/- 0.1 mm and 7.3%, and a centre to near-electrode ratio of -0.29 to -0.15 against 36 to 55. Under the adopted reading alone the block would be data_missing (no 1-burst well); metadata_insufficient is kept because the labels, which decide whether such a well exists, cannot be read with certainty from the archive.

Findings about the data

What attempts found about the dataset itself, for its owner and the archive.

Data missingThe dandiset holds CT26 monolayers at two packet rates (1 Hz and 0.2 MHz) and CHO monolayers with one burst label. It has no images at 1 kHz or 10 kHz, no 300- or 900-ns pulses, no thermochromic film images and no CHO burst series, so 9 of the paper's 20 findings cannot be tested for missing data. Two more are blocked for missing metadata: no file gives the electrode elevation, so wells with raised electrodes, if any are in the dandiset, cannot be identified, and the burst labels cannot be matched to 1 burst with certainty.
Metadata errorThe dandiset description and every file's protocol text give a 3.8-mm electrode spacing, a 3.2-kV first pulse and 75, 50 and 25% steps. The paper used a 15.3-mm array, a 7.2-kV first pulse and 16-pulse packets falling by 12.5% per phase, and the CT26 images show electrode footprints at the paper's spacing (best match 6.1 mm from the centre along each axis, against 6.4 mm from the paper's geometry, over 52 two-well areas). The voltage cannot be checked from the images; if the CT26 wells were exposed at 3.2 kV, they are not the paper's 7.2-kV experiments.
Description missingThe number of bursts is given only in the subject label (x1, x2, x3, x5); all but one CT26 subject say "Unknown repetitions", and the protocol text says only that the protocol "was either repeated 2, 3, or 5 times", without saying whether the first application counts, so the labels cannot be matched to the paper's "1 burst (x4)" to "6 bursts (x4)" with certainty. The CT26 files state no pixel size (the `resolution` attribute of the image data is 96 in CT26 files and 0.1 in CHO files, neither a pixel size), and no file gives the electrode elevation.
OtherThe CT26 images are on two scales without saying so. Whole-well images (21870 x 21289 pixels) fit the 35-mm well at 1.625 um per pixel, and two-well images fit the electrode spacing at 3.25 um per pixel (best match 4% below the paper's geometry). At 1.625 um the whole-well images match the electrode spacing 9% above the paper's geometry (3 areas), which was not resolved. Whole-well images also have higher backgrounds (mean of the four footprints 1061 to 4560 a.u., mean 2628, 7 YP areas, against 180 to 1121, mean 332, 45 YP areas, in two-well images), so their arbitrary units differ.
OtherThe paper's 0.2 MHz CANCAN result, a sharp central spot with 40-fold suppression near the electrodes, is not seen in the dandiset. At x1, read here as the paper's 2 bursts (x4), the 0.2 MHz CANCAN wells are unstained everywhere (group-mean profile within 32 a.u. of background); at x3 and x5 they are broadly stained, with a centre to near-electrode ratio of 1.4 to 4.0. The dandiset's images may be other experiments than those of the paper's figures.
Selection not documentedThe dandiset holds more wells per condition than the paper reports (for example 6 CT26 1 Hz CANCAN wells at label x1 where the paper's S1 has n = 3 per rate), and a CHO set at 1 Hz that the paper does not report. Which wells the paper used is not stated. No well was found deposited twice (peak normalised cross-correlation at most 0.039 over all 1654 pairs of areas in the same channel, whatever their condition and label, against 0.70 for a shifted, noisy copy of one area).
Metadata errorThe CHO files use session_start_time 2024-12-11, the conversion date, while their OME metadata give acquisition on 2024-08-14.
Methods not availableQuestions for the paper's authors. Does the subject label (x1, x2, x3, x5) count the repeats after the first burst or all bursts, and which wells make up the 2 bursts (x4) sets of Figures 3 to 5? Were the CT26 wells exposed at 7.2 kV, as in the paper, or 3.2 kV, as in the files? Were the wells imaged for propidium on May 24 2024 exposed with the electrodes touching the monolayer? In which electrode footprint were the 3 background ROIs placed, and how was the array centre located for the ROIs? When two error bars differ, which length must the gap exceed (Methods 2.5)? This attempt read the labels as repeats after the first burst, assumed the propidium wells had the electrodes touching, used the mean of the four footprints for background, located the centre from the footprints in the YP image, and required the gap to exceed the longer error bar.

Notebooks

Every registered entry that uses this dataset. An entry can hold several notebooks, and can use more than one dataset.

Reproduction Remote focusing of electroporation by MHz pulse packets (Silkuniene et al. 2025)

Tests the findings of Silkuniene, Silkunas, Pakhomova, Semenov and Pakhomov (2025) on dandiset 001047, which holds stitched YO-PRO-1, Hoechst and propidium images of CT26 monolayers exposed to CANCAN and control packets at 1 Hz and 0.2 MHz, and of CHO monolayers. It measures YP and propidium profiles along the diagonals of the electrode array and compares rates, protocols and dyes. The paper's 1 kHz and 10 kHz rates, pulse durations, electrode elevations, film thermometry and CHO burst series are not in the dandiset. All files

Reproducing Silkuniene et al. (2025) from dandiset 001047
reproduce_silkuniene_2025.ipynbExecuted, with outputs · Source

Verification

Round 1
2026-11-01
Agreed with 17 of 20 outcomes. Independent agent run by CatalystNeuro (same harness, separate context, did not write the notebook). Reviewed: Round 1. The executed notebook of commit 64d5a452b (author-answers) with the conclusion cells and the summary removed, the helper module and the claims file, with the paper. The verifier also read the archive's asset metadata and tested the helper functions on synthetic input. The reason the verifier gave is in its own words; agrees compares its outcome with the record it reviewed.
- fig3-tail-deviation-falls-with-rate: the verifier’s outcome was “Not reproduced”. Reversed at x5. The x1 "holds" measures deviation from a 0.2 MHz fit stuck at the width bound with negative R. With two rates the Spearman coefficient is +/-1 by construction, and the Lorentzian fits it depends on failed.
- fig4b-centre-no-rate-effect: the verifier’s outcome was “Partially reproduced”. The one available comparison has p = 0.93 (x1) and 0.54 (x5). The 1 kHz and 10 kHz comparisons are missing. At x1 both rates show no centre uptake at all (1.0 and -1.4 a.u.), so "no effect" there carries little information.
- fig7-sharpest-one-burst: the verifier’s outcome was “Not reproduced”. This claim is CT26 only, so the notebook's reason for blocking it (no CHO series) does not apply. If x1 means one burst, centre YP is -1.4 +/- 6.3 and mean FWHM 55.7 mm (2 of 4 fits at the bound), far from 1.1 mm and a ratio of 36 to 55. If x1 means two bursts, the condition is absent and the claim is blocked (metadata_insufficient).
Round 2
2026-11-01Agreed with 20 of 20 outcomes. Independent agent run by CatalystNeuro (same harness, separate context, did not write the notebook). Reviewed: Round 2. The full executed notebook of commit f282af7b2 (author-revise-1), the helper module, the claims file and the attempt record, with the paper. The verifier agreed with all 20 outcomes. It found four places where the text did not match the outputs: the data_missing finding said the dandiset has no raised electrodes and counted 11 data_missing blocks, two of which are not; the background of whole-well images was described without an output; two more areas (panels 62 and 63) look misplaced; and the duplicate check compared only areas of the same condition. In reply the two areas are flagged and the x1 claims are shown without them (no outcome changes; the fig3a control part then holds), the background is printed by image type, the duplicate check now covers all 1654 pairs in the same channel, and the findings are corrected. No outcome change is proposed.

Method differences

How the tests differed from the paper’s method or data, and the claims each applies to.

The error-bar rule (F8, Methods 2.5, "A gap exceeding the length of the error bars") is read as a gap beyond the longer s.e. arm, so that the gap exceeds each error bar (assumption); the mean-arm and shorter-arm readings are shown.fig3a-centre-enhanced fig3bc-rate-suppresses-off-target fig4b-near-electrode-decline
Background is the mean over the four footprints of 3 ROIs each (Methods 2.5 says 3 ROIs in one footprint without saying which); the darkest footprint is shown as a variant for the ratio claim.fig3a-centre-enhanced fig4b-40-fold-ratio
Burst labels read as repeats after the first burst, so that x1 is the paper's 2 bursts (x4) at both rates (assumption). The files' protocol text says the protocol "was either repeated 2, 3, or 5 times" upon completion, and under this reading x5 is 6 bursts (x4), the paper's largest number (Figure 7). Under the other reading (the label is the total) 2 bursts exists only at 1 Hz (x2); that reading is shown, with 1 Hz x2 against 0.2 MHz x1 and x3, which bracket 2 bursts. x5 is shown as a variant and decides nothing.fig3bc-rate-suppresses-off-target fig4b-40-fold-ratio
Only 1 Hz and 0.2 MHz are in the dandiset; the paper's other rates (1 kHz, 10 kHz and one more below 10 kHz) are not.fig3a-centre-enhanced
Burst labels read as repeats after the first burst, so that x1 is the paper's 2 bursts (x4) at both rates (assumption). The files' protocol text says the protocol "was either repeated 2, 3, or 5 times" upon completion, and under this reading x5 is 6 bursts (x4), the paper's largest number (Figure 7). Under the other reading (the label is the total) 2 bursts exists only at 1 Hz (x2); that reading is shown, with 1 Hz x2 against 0.2 MHz x1 and x3, which bracket 2 bursts. x5 is shown as a variant and decides nothing. The paper's controls are of unstated rate; the controls of both rates at the same label are pooled, and each rate is shown alone as a variant.fig3a-centre-enhanced
The array centre is found by matching the paper's electrode geometry (Methods 2.2) to the footprints in the YP image, or, where the match is weak, from the centroid of the stain (an assumption of four-fold symmetry); the paper placed ROIs by hand. Six areas whose centre is visibly misplaced are flagged and kept, and the claims they enter are also shown without them, with the same outcomes. Three areas whose centre fell near an image edge are dropped.fig3a-centre-enhanced
The parallel profile is measured on every area; the paper measured it in one set of experiments, which the dandiset does not identify.fig3bc-rate-suppresses-off-target
Only 1 Hz and 0.2 MHz. Burst labels read as repeats after the first burst, so that x1 is the paper's 2 bursts (x4) at both rates (assumption). The files' protocol text says the protocol "was either repeated 2, 3, or 5 times" upon completion, and under this reading x5 is 6 bursts (x4), the paper's largest number (Figure 7). Under the other reading (the label is the total) 2 bursts exists only at 1 Hz (x2); that reading is shown, with 1 Hz x2 against 0.2 MHz x1 and x3, which bracket 2 bursts. x5 is shown as a variant and decides nothing.fig3-lorentzian-profile
Only 1 Hz and 0.2 MHz, so a trend over rates reduces to one comparison. Burst labels read as repeats after the first burst, so that x1 is the paper's 2 bursts (x4) at both rates (assumption). The files' protocol text says the protocol "was either repeated 2, 3, or 5 times" upon completion, and under this reading x5 is 6 bursts (x4), the paper's largest number (Figure 7). Under the other reading (the label is the total) 2 bursts exists only at 1 Hz (x2); that reading is shown, with 1 Hz x2 against 0.2 MHz x1 and x3, which bracket 2 bursts. x5 is shown as a variant and decides nothing.fig3-tail-deviation-falls-with-rate
A fit whose width runs to its bound is taken as having no peak, so that no deviation from a bell-shaped fit can be measured; the order of the deviations alone is also shown.fig3-tail-deviation-falls-with-rate
Two rates, so one comparison under both readings of the claims file. Burst labels read as repeats after the first burst, so that x1 is the paper's 2 bursts (x4) at both rates (assumption). The files' protocol text says the protocol "was either repeated 2, 3, or 5 times" upon completion, and under this reading x5 is 6 bursts (x4), the paper's largest number (Figure 7). Under the other reading (the label is the total) 2 bursts exists only at 1 Hz (x2); that reading is shown, with 1 Hz x2 against 0.2 MHz x1 and x3, which bracket 2 bursts. x5 is shown as a variant and decides nothing.fig4b-centre-no-rate-effect
With two rates the power function passes through both means. Burst labels read as repeats after the first burst, so that x1 is the paper's 2 bursts (x4) at both rates (assumption). The files' protocol text says the protocol "was either repeated 2, 3, or 5 times" upon completion, and under this reading x5 is 6 bursts (x4), the paper's largest number (Figure 7). Under the other reading (the label is the total) 2 bursts exists only at 1 Hz (x2); that reading is shown, with 1 Hz x2 against 0.2 MHz x1 and x3, which bracket 2 bursts. x5 is shown as a variant and decides nothing.fig4b-near-electrode-decline
The dandiset gives no elevation. Assumed that the wells imaged for Pr had the electrodes touching the monolayer, the paper's default setting (Methods 2.2, "positioned ... until touching the bottom of the culture dish"). The data cannot check this.fig5ab-yp-spot-slightly-larger-than-pr
Burst labels read as repeats after the first burst, so that x1 is the paper's 2 bursts (x4) at both rates (assumption). The files' protocol text says the protocol "was either repeated 2, 3, or 5 times" upon completion, and under this reading x5 is 6 bursts (x4), the paper's largest number (Figure 7). Under the other reading (the label is the total) 2 bursts exists only at 1 Hz (x2); that reading is shown, with 1 Hz x2 against 0.2 MHz x1 and x3, which bracket 2 bursts. x5 is shown as a variant and decides nothing. The x3, x5 and pooled wells are shown; pooling mixes burst numbers whose centre YP differs over 1000-fold and is not the S2 condition.fig5ab-yp-spot-slightly-larger-than-pr
One x1 Pr area whose centre fell at the image edge is dropped. Pr areas are located in their own images; two of the five usable ones (one by the footprint match, one by the centroid) have visibly misplaced centres and are flagged, and results are shown without them.fig5ab-yp-spot-slightly-larger-than-pr
Same wells and assumptions as fig5ab-yp-spot-slightly-larger-than-pr.fig5b-gaussian-pr-lorentzian-yp

Provenance

Attempt record: 001047-silkuniene-2025-01
Paper: MHz compression of pulse packets facilitates remote focusing of electroporation
Date: 2026-11-01
Made by: An agent, run by CatalystNeuro
Dataset: DANDI:001047
Version: 0.250701.1609 (doi:10.48324/dandi.001047/0.250701.1609)
Run on: GitHub Actions
Harness: Claude Code
Model: claude-opus-5-5
Protocol: pilot-13

Transcripts of the agent runs

Each run’s prompt, messages and tool calls, with long output cut and the text of the paper left out, and its final report.

Other datasets that name the same paper

← DANDI:001038 DANDI:001048 →