LIMBR package

Submodules

LIMBR.imputable module

LIMBR.old_fashioned module

class LIMBR.old_fashioned.old_fashioned(filename, data_type, pool=None)

Bases: object

Performs a standard normalization procedure without SVD as a baseline.

This class performs simple quantile normalization and row scaling along with pool normalization for proteomics experiments using the same methods and interface employed in the sva class. This provides a baseline comparison point for data processed with LIMBR.

Parameters:

• filename (str) – Path to the input dataset.
• data_type (str) – Type of dataset, one of ‘p’ or ‘r’. ‘p’ indicates proteomic with two index columns specifying peptide and protein. ‘r’ indicates RNAseq with one index column indicating gene.
• pool (str) – Path to file containing pooled control design for experiment in the case of data_type = ‘p’. This should be a pickled dictionary with the keys being column headers corresponding to each sample and the values being the corresponding pooled control number.

Variables:

• raw_data (dataframe) – This is where the input data is stored.
• data_type (str) – This is where the data type (‘p’ or ‘r’) is stored.
• norm_map (dict) – This is where the assignment of pooled controls to samples are stored if data_type = ‘p’.

normalize(outname)

Groups peptides by protein and outputs final processed dataset.

These final results are then written to an output file.

Parameters:outname (str) – Path to desired output file.

pool_normalize()

Preprocessing normalization.

Performs pool normalization on an sva object using the raw_data and norm_map if pooled controls were used. Quantile normalization of each column and scaling of each row are then performed.

Variables:

• scaler (sklearn.preprocessing.StandardScaler()) – A fitted scaler from the sklearn preprocessing module.
• data_pnorm (dataframe) – Pool normalized data.

LIMBR.sva module

Module contents