deimos

deimos.build_factors(data, dims='detect')

Determine sorted unique elements (factors) for each dimension in data.

Parameters:

• data (DataFrame) – Feature coordinates and intensities.

• dims (str or list) – Dimensions to determine factors for. Attempts to autodetect by default.

Returns:

Unique sorted values per dimension.

Return type:

dict of array

deimos.build_index(data, factors)

Construct data index from precomputed factors.

Parameters:

• data (DataFrame) – Feature coordinates and intensities.

• factors (dict) – Per-dimension arrays of unique values.

Returns:

Index per dimension.

Return type:

dict of array

deimos.collapse(features, keep=['mz', 'drift_time', 'retention_time'], how=<function sum>)

Collpases input data such that only specified dimensions remain, according to the supplied aggregation function.

Parameters:

• features (DataFrame) – Input feature coordinates and intensities.

• keep (str or list) – Dimensions to keep during collapse operation.

• how (function or str) – Aggregation function for collapse operation.

Returns:

Collapsed feature coordinates and aggregated intensities.

Return type:

DataFrame

deimos.get_accessions(path)

Determines accession fields available in the mzML file.

Parameters:

path (str) – Path to mzML file.

Returns:

Dictionary of accession fields.

Return type:

dict of str

deimos.load(path, key='ms1', columns=None, chunksize=10000000.0, meta=None, accession={}, dtype=<class 'numpy.float32'>)

Loads data from HDF5 or mzML file.

Parameters:

• path (str or list of str) – Path to input file (or files if HDF5).

• key (str) – Access this level (group) of the HDF5 container. E.g., “ms1” or “ms2” for MS levels 1 or 2, respectively. HDF5 format only.

• columns (list) – A list of columns names to return. HDF5 format only.

• chunksize (int) – Dask partition chunksize. HDF5 format only. Unused when loading single file.

• meta (dict) – Dictionary of meta data per path. HDF5 format only. Unused when loading single file.

• accession (dict) – Key-value pairs signaling which features to parse for in the mzML file. mzML format only. See get_accessions() to obtain available values.

• dtype (data type) – Data type to encode values. mzML format only.

Returns:

Feature coordinates and intensities for the specified level. Pandas is used when loading a single file, Dask for multiple files. Loading an mzML file returns a dictionary with keys per MS level.

Return type:

DataFrame or dict of DataFrame

deimos.locate(features, by=['mz', 'drift_time', 'retention_time'], loc=[0, 0, 0], tol=[0, 0, 0], return_index=False)

Given a coordinate and tolerances, return a subset of the data.

Parameters:

• features (DataFrame) – Input feature coordinates and intensities.

• by (str or list) – Dimension(s) by which to subset the data.

• loc (float or list) – Coordinate location.

• tol (float or list) – Tolerance in each dimension.

• return_index (bool) – Return boolean index of subset if True.

Returns:

• DataFrame – Subset of feature coordinates and intensities.

• array – If return_index is True, boolean index of subset elements, i.e. features[index] = subset.

deimos.locate_asym(features, by=['mz', 'drift_time', 'retention_time'], loc=[0, 0, 0], low=[0, 0, 0], high=[0, 0, 0], relative=[False, False, False], return_index=False)

Given a coordinate and asymmetrical tolerances, return a subset of the data.

Parameters:

• features (DataFrame) – Input feature coordinates and intensities.

• by (str or list) – Dimension(s) by which to subset the data.

• loc (float or list) – Coordinate location.

• low (float or list) – Lower tolerance(s) in each dimension.

• high (float or list) – Upper tolerance(s) in each dimension.

• relative (bool or list) – Whether to use relative or absolute tolerance per dimension.

• return_index (bool) – Return boolean index of subset if True.

Returns:

• DataFrame – Subset of feature coordinates and intensities.

• array – If return_index is True, boolean index of subset elements, i.e. features[index] = subset.

deimos.multi_sample_partition(features, split_on='mz', size=500, tol=2.5e-05)

Partitions data along a given dimension. For use with features across multiple samples, e.g. in alignment.

Parameters:

• features (DataFrame or DataFrame) – Input feature coordinates and intensities.

• split_on (str) – Dimension to partition the data.

• size (int) – Target partition size.

• tol (float) – Largest allowed distance between unique split_on observations.

Returns:

A generator object that will lazily build and return each partition.

Return type:

Partitions

deimos.partition(features, split_on='mz', size=1000, overlap=0.05)

Partitions data along a given dimension.

Parameters:

• features (DataFrame) – Input feature coordinates and intensities.

• split_on (str) – Dimension to partition the data.

• size (int) – Target partition size.

• overlap (float) – Amount of overlap between partitions to ameliorate edge effects.

Returns:

A generator object that will lazily build and return each partition.

Return type:

Partitions

deimos.save(path, data, key='ms1', **kwargs)

Saves DataFrame to HDF5 or MGF container.

Parameters:

• path (str) – Path to output file.

• data (DataFrame) – Feature coordinates and intensities to be saved. Precursor m/z and intensities should be paired to MS2 spectra for MGF format.

• key (str) – Save to this level (group) of the HDF5 container. E.g., “ms1” or “ms2” for MS levels 1 or 2, respectively. HDF5 format only.

• kwargs – Keyword arguments exposed by to_hdf() or save_mgf().

deimos.slice(features, by=['mz', 'drift_time', 'retention_time'], low=[0, 0, 0], high=[0, 0, 0], return_index=False)

Given a feature coordinate and bounds, return a subset of the data.

Parameters:

• features (DataFrame) – Input feature coordinates and intensities.

• by (str or list) – Dimensions(s) by which to subset the data

• low (float or list) – Lower bound(s) in each dimension.

• high (float or list) – Upper bound(s) in each dimension.

• return_index (bool) – Return boolean index of subset if True.

Returns:

• DataFrame – Subset of feature coordinates and intensities.

• array – If return_index is True, boolean index of subset elements, i.e. features[index] = subset.

deimos.threshold(features, by='intensity', threshold=0)

Thresholds input DataFrame using by keyword, greater than value passed to threshold.

Parameters:

• features (DataFrame) – Input feature coordinates and intensities.

• by (str) – Variable to threshold by.

• threshold (float) – Threshold value.

Returns:

Thresholded feature coordinates.

Return type:

DataFrame