epi2me-labs/wf-single-cell: v3.3.1

Added

• Support for 10x Visium HD 3′ data, including the generation of 2 µm unbinned and 8 µm binned expression matrices. For details on how to preprocess this data, see percula. Currently the workflow supports only a single sample when processing such data.

Changed

• Calculation of sequencing saturation metrics has been overhauled in order to allow scaling to large datasets. The new method is more memory efficient. The gene and UMI saturation graphs are now presented in terms of numbers of UMIs rather than number of reads.
• Metrics in report have been renamed for clarity with their descriptions updated to better reflect their meaning.
• Two workflow steps (process_matrix and assign_features) have been amended to be more memory and time efficient, allowing for processing of larger datasets.
• Updated to align software with wf-template v5.6.2. This change does not have any functional impact on the workflow, but it ensures that the workflow remains compatible with the latest standards and practices.

Fixed

• Reads with unexpected UMI lengths were not removed in a consistent manner. This could lead to unexpected behaviour in several edgecases. The read removal is now more systematically handled across the dataset.