-
Notifications
You must be signed in to change notification settings - Fork 1
Terminology
The Abstract Beamline Interface makes use of some terms that are in common use in MX. Those terms have, in various contexts, been used with subtly different meanings. In some cases, different terms have also been used synonymously. We attempt a rigorous and canonical definition of those terms here, as they are used by the Abstract Beamline Interface.
A strategy is a partial specification of how a set of images are to be collected, and how those images are intended to be processed. In general, a strategy originates from a specific scientific perspective, and will typically be incomplete with respect to the detailed information needed to actually collect and process the images. Different types of strategy are incomplete in different ways.
The Abstract Beamline Interface currently defines just one type of strategy: the Geometric Strategy. A Geometric Strategy is a set of one or more sweeps (see next section) without any specification of (for example) their chronological ordering or subdivision. See the page on Strategy, Data Collection and Interleaving for more discussion.
Deciding on the strategy to use for data collection from a particular sample may involve the use of information about the sample that is already known. This information can include one or both of:
-
prior knowledge of the protein and crystal symmetry in general
-
results of processing and analysing the images from earlier data collections on the same or other samples
For a goniostat and detector of well-defined geometry, a sweep is defined as all of the following:
-
The scan axis of the goniostat (always Omega for a mini kappa)
-
A starting and ending setting of the scan axis; the sweep includes all settings of the scan axis between them.
-
For multi-axis goniostats, the settings of the goniostat axes other than the scan axis
-
When a crystal is mounted on the goniostat, these settings also define the crystal’s orientation
-
-
Wavelength
-
Detector position (distance, 2-Theta)
When applied to a mounted crystal whose orientation and unit cell parameters are known, a sweep implicitly defines a subset of the reflections that are, in principle, measurable from the crystal. (Some of these reflections may not be measurable in practice, due to such factors as crystal quality and sensitivity to radiation damage.)
During the derivation of an interleaved data collection from a set of sweeps, the sweeps are divided into smaller segments known as wedges. A wedge can be defined with reference to a sweep by specifying the starting setting of the scan axis, together with either the ending setting of the scan axis or the width of the wedge (such that the wedge is contained completely within its parent sweep).
The Abstract Beamline Interface does not have a class corresponding to
a wedge, but the attribute SampleCentred::wedgeWidth allows the
specification of a common wedge width in terms of image width and
number of images per wedge. This wedge width can then be used to
decompose a set of sweeps into wedges of constant width using this
value.
A scan is the action of collecting a set of diffraction images without any intervening discontinuity such as:
-
re-orientation of the sample
-
wavelength change
-
change of detector position
-
Omega → Omega ±180° transitions in an inverse-beam collection.
It includes or references all the parameters that are required to collect and process the images, including some that do not necessarily feature in the definitions of sweeps or wedges, in particular:
-
image width (depending on how the wedge width was defined)
-
image exposure time
-
beam intensity or attenuation
-
image filename template and image number of the first image in the scan.
In the Abstract Beamline Interface, a scan is defined with reference to a sweep; the purely geometric data collection parameters that form part of the definition of a sweep are not duplicated in order to define a scan.
With the above definitions, we can define a data collection as an uninterrupted, ordered sequence of scans. The significance of "uninterrupted" in this context is that a beamline is able to execute the entire data collection without a pause for any form of data analysis or decision-making about how to continue the experiment.
At first sight wedges and scans may seem very similar. The essential difference is that a scan defines the collection of an actual or potential set of images, whereas a wedge is a more abstract subdivision of a geometric strategy. Just like a sweep or geometric strategy, a wedge is incomplete with respect to the information needed to collect or process images.
During the derivation of a data collection from a geometric strategy, wedges are applied to sweeps to generate scans. Scans and wedges are not always strictly equivalent, however:
-
if the specification of wedge width is not an integral divisor of the width of the parent sweep, the scan that corresponds to the sweep’s final wedge will be narrower than the wedge width.
-
if two or more wedges adjoin, then they will be coalesced into a single scan