I’m using the metabolomics tool in Galaxy during my diploma thesis for direct infusion MS data. I hope, anyone can help me with my sample filtering problem.
In the sample filter of the metabolomics tool, it is possible to apply the sample filter within each sample class or not and I can set the minimum fraction to any value between 0 and 1. According todescription of Galaxy, the tool checks, whether greater-than the user-defined “minimum fraction” of samples contain an intensity value for a specific mass spectral peak. The intensity will be set to zero, if less than this number samples was found.
I assumed, that more intensity values will be set to zero, the higher the minimum fraction is. I’ confused, because there are more missing values with a minimum fraction of 90 percent than with 66.6 percent. Actually, I’m loosing samples with the lower “minimum fraction”, because the percentage of missing value is higher than my settings in the missing value sample filter.
For my approach, it is not necessary to apply the sample filter within each sample class. I tried both options anyway, but the lower “minimum fraction” leads in both versions to a stricter filtering process.
Can anybody help me with my misunderstanding?
Thank you very much in advance.