10.1. Auxiliary Readers — MDAnalysis.auxiliary

The auxiliary submodule contains code for reading ‘auxiliary’ data from a trajectory and allowing alignment with trajectory timesteps. Additional methods in MDAnalysis.coordinates.base.ProtoReader allow auxiliary data to be added and read alongside a trajectory.

Auxiliary data are timeseries accompanying a trajectory not stored in the regular trajectory file to be read by the trajectory Reader. They may be stored internally (e.g. in an array) or read from a file. In general, auxiliary data is assumed to be time ordered and contain no duplicates.

10.1.1. Supported formats

Currently supported formats:



Extension (if file)






Produced by Gromacs during simulation or analysis. Reads full file on initialisation.




Alternate xvg file reader, reading each step from the file in turn for a lower memory footprint. XVGReader is the default reader for .xvg files.




Produced by Gromacs during simulation. Reads full file on initialisation

10.1.2. Auxiliary API

Auxiliary readers inherit from the base AuxReader. In stand-alone use they allow iteration over each step in a set of auxiliary data:

aux = MDAnalysis.auxiliary.XVG.XVGReader('auxdata.xvg')
for auxstep in aux:

To iterate over only certain sections of the auxiliary:

for auxstep in aux[100:200]:
    # only steps 100-200

Or to skip steps:

for auxstep in aux[100::10]:
    # every 10th step starting at 100

A base AuxFileReader is also provided, extending AuxReader with attributes/methods for when auxiliary data is to be read from a file by keeping the file open and reading steps one-at-time as needed.

A get_auxreader_for() function is available to return an appropriate AuxReader class for a supplied format or set of auxiliary data, guessing the format from the datatype/file extension:

auxreader = MDAnalysis.auxiliary.core.get_auxreader_for('auxdata.xvg')
# will return the default XVGReader class
auxreader = MDAnalysis.auxiliary.core.get_auxreader_for(format='XVG-F')
# will return the XVGFileReader class

To directly load an instance of the guessed auxiliary reader class given the supplied auxdata and any additional auxiliary reader options, the function auxreader() can be used:

aux = MDAnalysis.auxiliary.auxreader('auxdata.xvg', dt=2)
for auxstep in aux:
   do_something(auxstep) Auxiliaries and trajectories

For alignment with trajectory data, auxiliary readers provide methods to assign each auxiliary step to the nearest trajectory timestep, read all steps assigned to a trajectory timestep and calculate ‘representative’ value(s) of the auxiliary data for that timestep.

‘Assignment’ of auxiliary steps to trajectory timesteps is determined from the time of the auxiliary step, dt of the trajectory and time at the first frame of the trajectory (obtained through a Timestep instance from the trajectory), as:

frame = floor((time_at_step - time_at_frame_0 + dt/2)/dt)

If there are no auxiliary steps assigned to a given timestep (or none within cutoff, if set), the representative value(s) are set to np.nan. Adding an auxiliary to a trajectory

Auxiliary data may be added to a trajectory Reader through the add_auxiliary() method. Auxiliary data may be passed in as a AuxReader instance, or directly as e.g. a filename, in which case get_auxreader_for() is used to guess an appropriate reader. e.g.:

u = MDAnalysis.Universe(PDB, XTC)
u.trajectory.add_auxiliary('pullforce', './pull_force.xvg')

As the trajectory frame is updated, the auxiliary data will be read correspondingly, updated, and the representative auxiliary value(s) will be available as e.g. u.trajectory.ts.aux.pullforce. Iterating by an auxiliary

The trajectory ProtoReader methods next_as_aux() and iter_as_aux() allow for movement through only trajectory timesteps to which one or more steps that fall within cutoff from a given auxiliary have been assigned. This may be used to avoid representative values set to np.nan, particularly when auxiliary data is less frequent:

u.trajectory.add_auxiliary('low_f', 'low_freq_aux_data.xvg')
for ts in u.trajectory.iter_as_aux('low_f'):
    do_something(ts.aux.low_f) # do something with 'low_f' auxiliary data without
                               # worrying about having to deal with np.nan

If the auxiliary data are more frequent and the cutoff (if set) is sufficiently high, next_as_aux() and iter_as_aux() behave the same as the ProtoReader next() and __iter__().

In order to access auxiliary values at every individual step, including those outside the time range of the trajectory, iter_auxiliary() allows iteration over the auxiliary independent of the trajectory:

for auxstep in u.trajectory.iter_auxiliary('pullforce'):

To iterate over only a certain section of the auxiliary:

for auxstep in u.trajectory.iter_auxiliary('pullforce', start=100, step=10):
    # every 10th step from 100

The trajectory remains unchanged, and the auxiliary will be returned to the current timestep after iteration is complete. Accessing auxiliary attributes

To check the values of attributes of an added auxiliary, use get_aux_attribute(), e.g.:

pullf_dt = u.trajectory.get_aux_attribute('pullforce', 'dt')

If attributes are settable, they can be changed using set_aux_attribute(), e.g.:

u.trajectory.set_aux_attribute('pullforce', 'data_selector', [1])

The auxiliary may be renamed using set_aux_attribute with ‘auxname’, or more directly by using rename_aux():

u.trajectory.rename_aux('pullforce', 'pullf') Recreating auxiliaries

To recreate an auxiliary, the set of attributes necessary to replicate it can first be obtained with get_description(). The returned dictionary can then be passed to auxreader() to load a new copy of the original auxiliary reader:

description = aux.get_description()
reload_aux = MDAnalysis.auxiliary.auxreader(**description)

The ‘description’ of any or all the auxiliaries added to a trajectory can be obtained using get_aux_descriptions():

descriptions = u.trajectory.get_aux_descriptions()

Get descriptions for selected auxiliaries only:

descriptions = u.trajectory.get_aux_descriptions(['pullf', 'pullx'])

And to reload:

for descr in descriptions:
    new_u.new_trajectory.add_auxiliary(**descr) AuxStep class

An AuxStep instance holds the auxiliary data for the current step. It is updated whenever a new auxiliary step is read.

AuxStep classes are derived from the base class AuxStep. The appropriate AuxStep class for a given auxiliary reader is identified in the reader by the _Auxstep attribute. Attributes

The following are inherited from AuxStep:


Current auxiliary step (0-based).


All recorded data for the current step, as a numpy array.


Time of current auxiliary step, as a float (in ps). Determined from the _data if time selection is enabled and a valid time_selector provided; otherwise calculated using dt and initial_time.


Auxiliary values of interest for the current step, as a numpy array. Determined from _data id data selection is enabled at a valid data_selector is provided; otherwise set to _data.

The following are stored in AuxStep. The parent auxiliary reader has the equivalent attributes, though non-private (no _) (see AuxReader class below).


Change in time between auxiliary steps (in ps). If not specified, will attempt to determine from auxiliary data; otherwise defaults to 1 ps.


Time of first auxiliary step (in ps). If not specified, will attempt to determine from auxiliary data; otherwise defaults to 0 ps.


Selection key to get time from full set of auxiliary data read with each step(_data) (ignored if time selection is not enabled by the reader). Type depends on the auxiliary format - e.g. where data is stored in columns, time_selector may be an index of ‘time’ column. Default value is None, in which case step time is calculated from dt, initial_time and step.


Selection key(s) to get data of interest from full set of auxiliary data read with each step (_data) (ignored if data selection is not enabled by the reader). As for time_selector, type depends on the auxiliary format. If None (default value), _data is returned.


Boolean of whether dt is constant throughout auxiliary data. Default is True. Methods

The following methods are inherited from AuxStep:


Setup appropriate attributes based on kwargs.

To enabled time/data selection in a particular AusStep, the following must be provided:


Return, as a float, the time value indicated by key from _data (the full set of data read in from the current step). Raise ValueError if key is not a valid time selector for the auxiliary format.


Return, as a ndarray, the value(s) indicated by key (may be e.g. a list of multiple individual ‘keys’) from _data. Raise ValueError if key is not a valid data selector for the auxiliary format.

Depending on the format of the auxiliary and hence the format of data, it may be necessary to overload the following for a particular AuxStep:


Return a np.array in the same format as data with all values set to np.nan. Used as the representative auxiliary value for a trajectory when no auxiliary steps are assigned to the current frame. AuxReader class Registry

In order to facilitate guessing of appropriate AuxReaders, all AuxReaders should set as appropriate format attribute. For files, this will be the expected file extension (in all caps). Replicating auxiliaries

The names of the necessary attributes for replicating an auxiliary are stored in required_attrs, initially set in the base AuxReader. If a particular AuxReader introduces additional attributes required to reload an auxiliary, these should be added. Attributes

The following attributes are inherited from AuxReader:


Name under which auxiliary data will be stored in trajectory.


Method to use in calculating representative auxiliary value for a timestep. Default is ‘closest’.


Cutoff (in ps) for ignoring auxiliary steps when calculating representative value(s) for a timestep.


The AuxStep object to store data for the current step. The particular AuxStep used will depend on the auxiliary format.


Total number of auxiliary steps.


data from each auxiliary step assigned to the last-read trajectory timestep.


Representative value(s) of auxiliary data for last-read trajectory timestep.

The following are stored in auxstep as private attributes (with _) but may be accessed from the auxiliary reader; see the AuxStep class above.








AuxFileReader provides:


File object for the auxiliary file.

Each auxiliary reader subclass will additionally set:


Value of ‘auxdata’ used to load the auxiliary - e.g. path to the file containing the auxiliary data. Will be used when getting the ‘description’ of the auxiliary for recreating.

Each auxiliary reader class must also identify an appropriate AuxStep with the _Auxstep class attribute Methods

The following methods are inherited from AuxReader:


Setup appropriate attributes based on kwargs.


Number of steps in auxiliary data.


Advance to next step.


Allow iteration through each auxiliary step.


If step is a single index, move to that step and return the AuxStep; if a list or slice, return an iterator over the specified auxiliary steps. See examples in Auxiliary API above.


Reposition to first step.


Read auxiliary steps assigned to trajectory step ts, calculate and update the representative value in ts. Return ts.


Read auxiliary steps assigned to trajectory step ts and calculate representative value.


Move to the auxiliary step before the trajectory timestep ts, such that calling _read_next_step reads the first step assigned to ts or, when auxiliary data are less frequent and no steps are assigned to ts, the first step following ts.

step_to_frame(step, ts, return_time_diff=False)

Return the frame number of the trajectory described by ts to which the auxiliary step step is assigned. Optionally also return the difference it time between the step and the returned frame.


Return the time of the auxiliary step step.


Return the frame number of the next trajectory frame (after the current auxiliary time) for which a representative auxiliary value can be calculated (i.e., there is at least one assigned auxiliary step within cutoff).


Return the representative value calculated from data following the method specified by represent_ts_as and cutoff.


Entry method for Context Manager (returns self).


Exit method for Context Manager (calls close()).


Calls close().


Get the values of the attributes required for replicating an auxiliary (as listed in required_attrs) and return as a dictionary.


Check for equality by checking each of the attributes required for replicating an auxiliary (as listed in required_attrs) are equal.

Each AuxReader must subclass AuxReader and addionally define:

__init__(auxdata, **kwargs)

Additional processing of kwargs and any necessary initial processing of auxdata. Must super() through AuxReader.


Read the data from the next auxiliary step and update auxstep as appropriate. Raise StopIteration when attempting to read past last step.


Move to and read step i (0-based) from the auxiliary data. Raise ValueError when i is out of range. Update the auxstep and return.

Depending on the format of the auxiliary data, it may also be necessary to define/overwrite the following:


Return the list of times for each step (only required if constant_dt may be false).


Return the total number of steps (only required if _n_steps not otherwise set during __init__).


Reposition before the first step


For convenience, when reading auxiliary data from an open file, one step at a time, AuxFileReader extends AuxReader by providing the following (these may be overwritten by subclasses as appropriate):

__init__(auxfile, **kwargs)

Open auxfile and any additional processing of kwargs.


Seek to the start of auxfile


Close auxfile


Iterate through all steps until step i is reached.


Close auxfile and reopen.