AutoED multiplex

The xia2 multiplex command combines several successful integrated results into one. The AutoED multiplex uses only the results of a single pipeline (usually the default pipeline).

  • The multiplex processing is triggered every time the dedicated pipeline (e.g., the default) updates the AutoED automatic report. Triggering the multiplex does not necessarily mean the xia2 multiplex command will run.

  • The AutoED will first determine to which sample the triggering dataset belongs (for more on samples, see below) and check if the percentage of indexed spots is above a certain threshold (see multiplex_indexing_percent_threshold parameter in the global configuration file).

  • If the dataset is high quality (above the indexing threshold), the AutoED will add it to its list of samples to process with multiplex. The multiplex command runs on every fifth dataset added (this can be modified with the multiplex_run_on_every_nth parameter in the configuration file). Note that this does not mean multiplex will run on every fifth measured dataset but on every fifth high-quality dataset for a given sample.

Multiplex samples

Based on datasets processed so far, AutoED makes few assumptions about how diffraction files are structured. A typical path of a single dataset looks like this.

/../ED/SAMPLE_DIRS/CRYSTAL_DIR/SWEEP_DIR/DATA_master.h5

Here, ED is the data root directory (a directory where AutoED looks for diffraction images). The assumption is there is only one data root directory in a dataset path (a complete path of a master file). CRYSTAL_DIR is a directory where we keep diffraction data obtained from individual crystals. These directories usually start with a date (e.g., 20240521_nav22_...). A CRYSTAL_DIR can contain one or more SWEEP_DIR-s (e.g. sweep0, sweep1, sweep2, etc). The assumption is that everything between the ED directory and the CRYSTAL_DIR is a grouping of different samples.

For example, if we have the following structure

../ED/sample_01/20250215_nav15_124.543.22/sweep0/...master.h5
../ED/sample_02/20250215_nav11_172.255.33/sweep0/...master.h5

AutoED assumes we are dealing with two different samples. You can have as many directories for sample grouping as you like, so something like this.

ED/abc/123/sample_02/20250215_nav11_.../sweep0/...master.h5

is treated as a sample.

The structure of the sample directories is copied into the multiplex directory. In the case when crystal directories are placed directly in the ED data root

../ED/20250215_nav7.../sweep0/...master.h5
../ED/20250216_nav5.../sweep0/...master.h5

These crystals would be assigned to the default sample, and the output of their processing would be in multiplex/default_sample.