Input Formats

All omics data is tabular. The rows are the samples, and the columns are the features. The rows and columns can be annotated with metadata. The data is stored in a dataframe object, which is a 2D labeled data structure with columns of potentially different types.

Features:

• df.index.name - name of the index (row labels)

• df.columns.name - name of the columns (column labels)

MultiIndex objects are not supported as most libarires do not support them. metadata is stored in a separate dataframe.

The approach is to keep it simple, just use a dataframe object, which can be easily serialized to a human readable format, e.g. a csv file. The file-format can be anything and the validity of the data is check when data is used, mainly using pandera or a few custom functions.