biom.table.Table.transform

Table.transform(f, axis='sample', inplace=True)

Iterate over axis, applying a function f to each vector.

Only non null values can be modified and the density of the table can’t increase. However, zeroing values is fine.

Parameters:

ffunction(data, id, metadata) -> new data

A function that takes three values: an array of nonzero values corresponding to each observation or sample, an observation or sample id, and an observation or sample metadata entry. It must return an array of transformed values that replace the original values.

axis{‘sample’, ‘observation’}, optional

The axis to operate on. Can be “sample” or “observation”.

inplacebool, optional

Defaults to True. Whether to return a new table or modify itself.

Returns:

biom.Table

Returns itself if inplace, else returns a new transformed table.

Raises:

UnknownAxisError

If provided an unrecognized axis.

Examples

>>> import numpy as np
>>> from biom.table import Table

Create a 2x3 table

>>> data = np.asarray([[0, 0, 1], [1, 3, 42]])
>>> table = Table(data, ['O1', 'O2'], ['S1', 'S2', 'S3'],
...               [{'foo': 'bar'}, {'x': 'y'}], None)
>>> print(table) 
# Constructed from biom file
#OTU ID S1  S2  S3
O1  0.0 0.0 1.0
O2  1.0 3.0 42.0

Create a transform function

>>> f = lambda data, id_, md: data / 2

Transform to a new table on samples

>>> table2 = table.transform(f, 'sample', False)
>>> print(table2) 
# Constructed from biom file
#OTU ID S1  S2  S3
O1  0.0 0.0 0.5
O2  0.5 1.5 21.0

table hasn’t changed

>>> print(table) 
# Constructed from biom file
#OTU ID S1  S2  S3
O1  0.0 0.0 1.0
O2  1.0 3.0 42.0

Tranform in place on observations

>>> table3 = table.transform(f, 'observation', True)

table is different now

>>> print(table) 
# Constructed from biom file
#OTU ID S1  S2  S3
O1  0.0 0.0 0.5
O2  0.5 1.5 21.0

but the table returned (table3) is the same as table

>>> print(table3) 
# Constructed from biom file
#OTU ID S1  S2  S3
O1  0.0 0.0 0.5
O2  0.5 1.5 21.0