When I upload my data in a .csv file to metaboanalyst and try to conduct a One-way ANOVA Test, the following error message pops up:
My data has 70 samples and 56739 features and is 25,989 KB.
How can this error be resolved?
Thank you! Here are the updated details based on that resource your provided:
1). Tool and module used: I used the “One-way Analysis of Variance (ANOVA)” tool within the “Statistical Analysis [one factor]” module.
2). Here is a link to the data set: Metaboanalyst_data.csv - Google Drive
3). I selected the csv file with samples in columns (unpaired), didn’t filter the data, and didn’t normalize the data.
4). This post does not relate to using R packages.
5). These data are LC-MS untargeted analysis data for soil-water extracts. The data were exported from mzmine.
That is literally the same error/troubleshooting I posted about yesterday for pathway analysis (even using sample data).
I see that you don’t want to filter and normalization - simply would like to perform one-way ANOVA on your data containing 70 samples (12 groups) by 56738 peaks.
Such practice only aims to use the computing power of our server - which is not what it is designed for. Please refrain from doing this as the public server has limited capacity and is used by many researchers.
Two comments:
-
Your approach does not followed omics data processing practice - you need to perform filtering and normalization. In addition to computing issue, you will lose statisitcal power due to multi-testing punishment
-
Use PCA + linear model (for specific comparison) to gain more deep insights. Interpreting ANOVA on 12 samples will be hard and you need to perform post hoc analysis which is another issue