Automated Assays Automated Assays

Run

group_comparison_anova 0.1.0

Run against the fork of Abnormal expression of ATP1A1 and ATP1A2 in breast cancer . Run identifier 2d090760-f2ae-423e-849f-dc58da829431.

What the run did

54,673 features were tested across 141 samples in 5 groups, and 22,426 of them came out differential.

The comparison used 141 samples in 5 groups, out of 155 samples in 6 groups. Not part of the comparison: CellLine.

The evidence manifest

The cohort the run used

This run counted 141 experimental units in the paper's 5 groups (30 Her2, 41 TNBC, 11 Healthy, 29 Luminal A, 30 Luminal B), which is the same group count as the paper states. This is an agreement of counts: the paper publishes no sample identifiers, so it is not an identity of specimens, and the paper does not say how it handled its own technical replicates. The repository holds 178 arrays; 46 of them carry the repository's own replicate marker, and averaging them removed 23 arrays and left the experimental units this run counted. CellLine 14 is outside the paper's cohort.

  • — The paper publishes no sample identifiers, so what is compared here is the number of experimental units in each group and not the identity of each specimen.
  • — The paper does not publish how it handled its own technical replicates, so the rule that produced these units is the repository's structure as the platform read it and not the paper's own rule.
  • — The experimental units are the repository's arrays averaged after the depositor's own replicate marker was removed.

What the paper published, and what the run found

ATP1A1 was overexpressed approximately 1.5 times in all groups of breast cancer samples (p<0.05). Coincidently, ATP1A2 expression decreased by more than 2 times (p<0.05).

  • ATP1A1 partly reproduced 4 contrasts · 0 reproduced · 0 direction only · 3 partly · 1 contradicted
    Luminal A against Healthy partly reproduced
    244606_at contradicted paper 1.53 , p 0.009016 run 0.967619 , p 2.38e-6 36.8% from the published value , tolerance 25.0%
    220948_s_at reproduced paper 1.53 , p 0.009016 run 1.477288 , p 0.002553 3.4% from the published value , tolerance 25.0%
    Luminal B against Healthy contradicted
    244606_at contradicted paper 1.38 , p 0.04454 run 0.973736 , p 2.38e-6 29.4% from the published value , tolerance 25.0%
    220948_s_at significance only paper 1.38 , p 0.04454 run 1.365716 , p 0.002553 1.0% from the published value , tolerance 25.0%
    Her2 against Healthy partly reproduced
    244606_at contradicted paper 1.66 , p 0.005926 run 0.977794 , p 2.38e-6 41.1% from the published value , tolerance 25.0%
    220948_s_at reproduced paper 1.66 , p 0.005926 run 1.708049 , p 0.002553 2.9% from the published value , tolerance 25.0%
    TNBC against Healthy partly reproduced
    244606_at contradicted paper 1.44 , p 0.015725 run 0.976144 , p 2.38e-6 32.2% from the published value , tolerance 25.0%
    220948_s_at reproduced paper 1.44 , p 0.015725 run 1.621305 , p 0.002553 12.6% from the published value , tolerance 25.0%

    Luminal A: the observed change is not an increase; the run goes the other way (published ratio 1.5300, run ratio 0.9676); Luminal B: the observed change is not an increase; the run goes the other way (published ratio 1.3800, run ratio 0.9737); the run agrees with the published ratio 1.3800 but reaches a significance the paper did not (published p 0.044540, run p 0.096332 against the same two groups); Her2: the observed change is not an increase; the run goes the other way (published ratio 1.6600, run ratio 0.9778); TNBC: the observed change is not an increase; the run goes the other way (published ratio 1.4400, run ratio 0.9761)

  • ATP1A2 direction only 4 contrasts · 0 reproduced · 4 direction only · 0 partly · 0 contradicted
    Luminal A against Healthy direction only
    203296_s_at direction only paper 0.401606 , p 1.85e-7 run 0.036208 , p 1.22e-37 91.0% from the published value , tolerance 25.0%
    203295_s_at direction only paper 0.401606 , p 1.85e-7 run 0.224409 , p 3.09e-16 44.1% from the published value , tolerance 25.0%
    Luminal B against Healthy direction only
    203296_s_at direction only paper 0.396825 , p 8.5e-9 run 0.028244 , p 1.22e-37 92.9% from the published value , tolerance 25.0%
    203295_s_at direction only paper 0.396825 , p 8.5e-9 run 0.222551 , p 3.09e-16 43.9% from the published value , tolerance 25.0%
    Her2 against Healthy direction only
    203296_s_at direction only paper 0.359712 , p 5.48e-8 run 0.029065 , p 1.22e-37 91.9% from the published value , tolerance 25.0%
    203295_s_at direction only paper 0.359712 , p 5.48e-8 run 0.225171 , p 3.09e-16 37.4% from the published value , tolerance 25.0%
    TNBC against Healthy direction only
    203296_s_at direction only paper 0.348432 , p 2.08e-11 run 0.031254 , p 1.22e-37 91.0% from the published value , tolerance 25.0%
    203295_s_at direction only paper 0.348432 , p 2.08e-11 run 0.233499 , p 3.09e-16 33.0% from the published value , tolerance 25.0%

    Luminal A: the run gets the direction right but its ratio 0.0362 is 91.0% from the published ratio 0.4016, outside the declared tolerance of 25%; the run gets the direction right but its ratio 0.2244 is 44.1% from the published ratio 0.4016, outside the declared tolerance of 25%; Luminal B: the run gets the direction right but its ratio 0.0282 is 92.9% from the published ratio 0.3968, outside the declared tolerance of 25%; the run gets the direction right but its ratio 0.2226 is 43.9% from the published ratio 0.3968, outside the declared tolerance of 25%; Her2: the run gets the direction right but its ratio 0.0291 is 91.9% from the published ratio 0.3597, outside the declared tolerance of 25%; the run gets the direction right but its ratio 0.2252 is 37.4% from the published ratio 0.3597, outside the declared tolerance of 25%; TNBC: the run gets the direction right but its ratio 0.0313 is 91.0% from the published ratio 0.3484, outside the declared tolerance of 25%; the run gets the direction right but its ratio 0.2335 is 33.0% from the published ratio 0.3484, outside the declared tolerance of 25%

  • ATP1A3 reproduced 4 contrasts · 4 reproduced · 0 direction only · 0 partly · 0 contradicted
    Luminal A against Healthy reproduced
    214432_at reproduced paper 0.952381 , p 0.429089 run 0.987372 , p 0.002985 3.7% from the published value , tolerance 25.0%
    Luminal B against Healthy reproduced
    214432_at reproduced paper 1.03 , p 0.308298 run 0.991701 , p 0.002985 3.7% from the published value , tolerance 25.0%
    Her2 against Healthy reproduced
    214432_at reproduced paper 0.961538 , p 0.768041 run 0.995745 , p 0.002985 3.6% from the published value , tolerance 25.0%
    TNBC against Healthy reproduced
    214432_at reproduced paper 0.961538 , p 0.527878 run 0.994019 , p 0.002985 3.4% from the published value , tolerance 25.0%

3 published claims were checked.

The paper’s method, step by step

Method verdict: unresolved. A step marked as not implemented here is a step this platform does not run, which is why this work is a re-analysis of the deposited data and not a reproduction of the paper’s pipeline.

  1. Microarray preprocessing Not implemented here

    Preanalytical procedures consisted of a robust multichip analysis (RMA) algorithm 21, including background correction, probe set signal integration, and quantile normalization. For this purpose, we used Expression Console 1.4 software (Affymetrix, Inc. USA).

    Software named by the paper: Expression Console 1.4

    Operations this platform does not have: microarray_rma_normalise

  2. One-way analysis of variance Implemented here

    We utilized Transcriptome Analysis Console 3.0 software (Affymetrix, Inc. USA) to analyze the obtained CHP files and to detect differentially expressed genes using one-way between subjects ANOVA.

    Software named by the paper: Transcriptome Analysis Console 3.0

  3. Group comparison against a reference group Not implemented here

    NKA genes expression in breast cancer samples relative to normal breast tissue.

    Operations this platform does not have: group_relative_expression

The software the paper used, and how to get it

Expression Console 1.4 software and Transcriptome Analysis Console 3.0 software (Affymetrix, Inc. USA) are available after free customer registration at: http://www.affymetrix.com/support/technical/software_downloads.affx.

http://www.affymetrix.com/support/technical/software_downloads.affx

Which probe the platform placed each gene on

The paper's method text names the analysis but not the probe set, so the platform infers which probe carries each gene and records the inference here.

A probe that reproduces the published ratio within the band is a hypothesis about which probe the paper used. It is not independent confirmation that the paper measured the gene on that probe, because the published values are the thing the band was applied to.

  • ATP1A1 · 220948_s_at · sequence_tagged

    Probe 220948_s_at reproduces the published value in every group (4 of 4), and no other candidate does.

  • ATP1A2 · sequence_tagged

    No candidate probe reproduces the published value within the declared tolerance. The closest is 203295_s_at, whose ratios sit 39.6% from the published values on average.

  • ATP1A3 · 214432_at · legacy_clone

    The platform carries one probe for ATP1A3, so the published value is placed without a choice to make.