Input data must be submitted in a tab-separated values (TSV) file which is a plain text format for storing data in a tabular structure. It can be prepared with a help of a spreadsheet software but must be saved as a plain text format with a tabulation as a field separator. Although any file's extension is accepted (. tsv, .txt, .csv, etc.), only tabulation is accepted as a field separator.
Each input file may include data related to several analytes, but user must keep in mind that all analytes in one file will have models learned with the same parameters except when the model parameter 'k' is not fixed, and the model is selected based on the lowest BIC. If the nature of the analyte data requires models with essentially different parameters they'd be better split into several files.
Each column must have a name formatted according to one of the following patterns:
'NAME (test)': contains test samples (real numbers)
'NAME (ref)': contains reference samples (real numbers)
'NAME (query-name)': contains query samples which will be classified with a model learned on test sample (real numbers, optional).
'id (test)': contains identification for all test samples (characters, optional)
'id (ref)': contains identification for all reference samples (characters, optional)
'id (query-id)' contains identification for query samples located at the right of this column till the next 'id-query' column or the end of columns (characters, optional)
where italicized parts can be customized. 'NAME' represents the analyte name and must be the same in all columns related to a given analyte (for more info, see 'Workflow examples')