CLC Single Cell Analysis Module latest improvements - Bioinformatics Software

CLC Single Cell Analysis Module 22.1

Released on 04.04.2022

Improvements

Importers

A preview of how the cell barcode and sample will be interpreted has been added to all single cell importers.
MEX importers have been updated to
- Allow an optional header in the features.tsv file
- Allow an optional header in the barcodes.tsv file
- Allow multiple columns that are tab-separated in the barcodes.tsv file. The barcodes need to be present in the first column.
A new ‘Sample’ parameter for naming samples has been added to the Expression Matrix and Peak Count Matrix importers.

Train Cell Type Classifier

Train Cell Type Classifier has been updated to optionally use validation data to produce a report summarizing the performance of the produced classifier.
A new tutorial Train a cell type classifier for single cell RNA-Seq analysis is now available.

QIAGEN Cell Ontology

A new tool Browse QIAGEN Cell Ontology has been added that opens the QIAGEN Cell Ontology and allows browsing for specific cell types.
QIAGEN Cell Ontology has been updated with new cell types.
A new option ‘Map clusters to QIAGEN Cell Ontology’ has been added to the Expression Matrix importer in Loom format.
Clusters representing cell types that are present in the QIAGEN Cell Ontology can be clicked to open the ontology to the corresponding type.

DR plots

Dimensionality reduction and phase portrait plots have been improved with new side panel settings allowing:
- Removal of cells that are not selected or have no value
- Dot and arrow size adjustment
- Addition of labels and legends
Dimensionality reduction plot editors can be opened in a side-by-side view mode, allowing different selections and colorings in different editors.
Improved the default colors when overlaying Transcription Factor counts in a dimensionality reduction plot, such that differences can now be seen between groups of cells. Previously, most transcription factors gave an apparently uniform color to all cells, because the upper position of the color slider was determined by the highest value in the plot, and this was often much higher than the other values.

General

QC for Single Cell has been changed to support automatic calculation of parameters related to droplets ambient profiles.
A new export tool, Export TXT Expression Matrix, for export of raw counts in tabular formats has been added.
Single Cell RNA-Seq Analysis optionally outputs unmapped reads.
The Chromatin Accessibility and Expression Analysis from Matrix workflow has been updated for better usability when using the on-the-fly import functionality.
Reference data sets, Single Cell hg38 (Ensembl) and Single Cell Mouse (Ensembl), have been updated with new Cell Type Classifiers.
The output of Convert Clonotypes to Cell Annotations has been updated to include the ‘Clone size’.
Improved the runtime of Normalize Single Cell Data when multiple matrices are provided as input.
Split Read Mapping by Cell generates a report with descriptive statistics.
The Matrix Track view has been extended with an option to show feature labels.
Improved the support for working with more than 100K cells.
Right-clicking on a phase portrait plot no longer includes options to launch tools. These right-click options are still available in dimensionality reduction plots.
Various minor improvements.

Bug fixes

Fixed an issue where Create Subset in dimensionality reduction plot editors, and Extract to Table in dimensionality reduction and phase portrait plot editor,s would not take into account unsaved changes to clusters.
Fixed an issue where Import Expression Matrix in Loom format was failing to link cell types from the QIAGEN Cell Ontology found in clusters.
Fixed an issue where an error would be shown when a selection was made on an Expression Matrix or Peak Count Matrix in a Track List and a table view of the matrix was subsequently opened.
Fixed an issue with the dimensionality reduction plot editor that would sometimes fail when loading plots with more than 100K cells.
Fixed an issue where “Extract to Table” in the dimensionality reduction plot editor would:
- For spliced/unspliced matrix selections: put the expression in all three columns (expression, spliced, unspliced)
- For velocity matrix selections: put the spliced read count in all three columns (spliced, unspliced, velocity)

CLC Single Cell Analysis Module 22.0

Released on January 11, 2022

New features

RNA Velocity

RNA Velocity analysis is fully supported via RNA-Seq workflows and new tools:
- The following workflows have been updated to calculate velocity, if possible, and produce relevant outputs (phase portraits and a ranking of velocity genes):
- Tools:
  - Single Cell Velocity Analysis calculates velocities from expression matrices that contain spliced / unspliced counts.
  - Differential Velocity for Single Cell calculates genes for which velocity estimates are statistically different between groups of cells.
  - Score Velocity Genes scores velocity genes for the entire sample / group of cells to help identify which velocity genes drive the observed dynamics.
  - Create Phase Portrait Plot creates an interactive and browsable scatter plot.

Chromatin Accessibility

Single Cell Chromatin Accessibility analysis is fully supported via the following new workflows and tools:
- The following tools have been added to support import of 10x ATAC matrices:
- The following tools have been added to support export of 10x ATAC matrices:
  - 10x HDF5 Peak Count Matrix
  - MEX Peak Count Matrix

Optionally the nearby genes and transcription factors can be imported or exported.

Workflows:
- Chromatin Accessibility Analysis from Reads analyzes ATAC-Seq reads to produce a UMAP plot, annotated with clusters and linked to peak, nearby genes and transcription factors.
- Chromatin Accessibility Analysis and Expression Analysis from Reads jointly analyses 10x Multiome ATAC and GEX data to produce a UMAP plot, annotated with clusters and linked to peak, nearby genes and transcription factors. In addition, velocity analysis is performed.
- Chromatin Accessibility Analysis and Expression Analysis from Matrix jointly analyses 10x Multiome ATAC and GEX matrices to produce a UMAP plot, annotated with clusters and linked to peak, nearby genes and transcription factors.
Tools:
- Single Cell ATAC-Seq Analysis analyzes ATAC-Seq read mappings and produces a filtered and normalized Peak Count Matrix. The peak count matrix contains peaks for each cell, their nearby genes and their transcription factors.
- Split Read Mapping by Cell produces read mappings for groups of cells and Graph Tracks to aid visualization of ATAC-Seq peaks.
- Differential Accessibility for Single Cell calculates peaks / nearby genes / transcription factors for which counts are statistically different between groups of cells.
It is possible to use the following tools with Peak Count Matrices, either standalone or in combination with an Expression Matrix:

Analysis including both matrix types will only consider cells that are present in both matrices.

Annotate Reads with Cell and UMI has new options “10x Chromium Single Cell ATAC” and “10x Chromium Single Cell Multiome ATAC” for demultiplexing 10x ATAC data types.
- When ATAC multiome is selected, barcodes are translated to match ATAC GEX.

Improvements

MEX format Import

Import Expression Matrix in MEX Format has been extended with options to take separate matrix files for spliced and unspliced reads. When these are used, the expression file may be omitted and the expressions are calculated from these counts.
Import Expression Matrix in MEX Format (archive) has been extended with the same option. The two matrix files must be called “spliced.mtx” and “unspliced.mtx”. This matches the convention of STARsolo.

MEX Expression Matrix Export

The MEX Expression Matrix (Archive) exporter is now called MEX Expression Matrix.
The MEX Expression Matrix exporter has been extended with the ability to export spliced and unspliced counts. They will be in files with fixed names “spliced.mtx” and “unspliced.mtx” matching STARsolo convention.
The MEX Expression Matrix exporter now outputs the three files (barcode, features, matrix) in .tar format directly instead of wrapping them in an archive.

Improvements to UMAP and tSNE plot

The dimensionality reduction plot editor has been extended to have separate panels for expression, peak and velocity matrices.
Right-clicking on the plot provides fast access to the following tools:
Add Information to Plot has been extended to take velocity matrices

General

Single Cell RNA-Seq Analysis has been updated to output a new expression matrix which additionally contains counts for spliced and unspliced reads.
Single Cell TCR-seq Analysis has been updated to better handle situations where the J segment cannot be uniquely identified for a clonotype. This change is not expected to impact the results significantly but may lead to slight changes in read counts as well as identification of clonotypes with low read counts.
Text files containing tables with columns separated by tabs (.tsv) can now be imported using Import Cell Annotations and Import Cell Clusters.
The “Empty droplets filter” in the QC for Single Cell tool has been updated to more easily enable retaining as cells only the barcodes above the automatically detected knee in the log-log rank plot without performing simulation-based tests.
When importing expression matrix in Loom format appropriate attributes are provided for easy selection.
Annotate Reads with Cell and UMI has a new option “10x Chromium Single Cell 5′” with the same read structure as “10x Chromium Single Cell 3′ v2”.
Import Cell Clusters, Import Cell Annotations and Import Cell Clonotypes have been updated to accept any type of data matrix (rather than just an expression matrix) as a parameter for setting the sample name.
QIAGEN Cell Ontology has been updated with new cell types.
Various minor improvements.

Bug fixes

Fixed an issue with fitting Negative Binomial Generalized Linear Model. This leads to slightly different results from Normalize Single Cell Data.
Various minor bug fixes.

Changes

Re-organized the structure of the toolbox to accommodate the new features better.
The Single Cell Workflows can now be located in the Template Workflows section of the toolbox.
When using workflows from previous versions of the Single Cell Analysis Module major adjustments may be required. Workflows containing the following tools needs to be manually redrawn by removing and re-adding:

Additional redrawing of connections may be required

The Count intronic reads” option from Single Cell RNA-Seq Analysis has been renamed to “Include intronic reads in total expression”.

Single Cell Analysis Module 21.1

Released on 24.06.2021

New features

T Cell Repertoire analysis is fully supported via the following new workflows and tools:
- Workflows:
  - Immune Repertoire Analysis from Reads (10x VDJ): Analyzes 10x VDJ reads to produce clonotypes at the cell level.
  - Immune Repertoire and Expression Analysis from Reads (10xVDJ): Jointly analyzes 10x 5′ GEX and VDJ reads to produce a UMAP plot, fully populated with clusters, cell types, and clonotypes.
  - Immune Repertoire and Expression Analysis from Clonotypes and Matrix: Jointly analyzes imported clonotypes and expression matrices to produce a UMAP plot, fully populated with clusters, cell types, and clonotypes.
- Tools:
  - Import Cell Clonotypes in Cell Ranger Contig format or AIRR format.
  - Single Cell Immune Repertoire Analysis to produce cell clonotypes from reads by mapping to TCR reference segments.
  - Filter Cell Clonotypes to remove low quality or uninteresting clonotypes based on various criteria.
  - Combine Cell Clonotypes into a single larger Cell Clonotypes element to make analysis of multiple samples easier.
  - Compare Single Cell Immune Repertoires across samples or across different subsets of T cells.
  - Convert Clonotypes to Cell Annotations that can be used for coloring cells by clonotype in a tSNE or UMAP plot.
Doublet detection and removal is available from QC for Single Cell.
The Create Expression Plot tool can create Violin plots.
Create Heat Map for Cell Abundance takes cell clusters and/or cell annotations. Based on two selected categories it then produces a heat map with one category as rows and one as columns. The value in each heat map entry is the percentage of cells in the combination reflected by the entry. The tool can also be launched from a UMAP or tSNE plot.
Update to Common Sample Name has been introduced for changing the name of one sample to be the same as an accompanying analyte sample.

Improvements

Import

It is possible to use Import Cell Annotations and Import Cell Clusters when selecting input to a workflow, so as to perform on-the-fly import.
A new “Cell format” option in Import Cell Annotations and Import Cell Clusters allows the extraction of sample and barcode information from one column in the imported file. It also allows text, such as a constant prefix, to be trimmed from all barcodes.
When importing an expression matrix, a feature with a versioned id (e.g. “ENSG00000243485.5”) will be matched to the genomic coordinates of a feature in a track with no version information (e.g. “ENSG00000243485”).
The Import Expression Matrix tools no longer assign genomic coordinates to imported features when there is ambiguity. For example, if no features have ids, and a feature with the same name is present on two chromosomes, then neither set of genomic coordinates will be assigned to an imported feature with that name.
Import Expression Matrix in Loom Format imports more of the data present in a Loom file as cell annotations.

Cell type prediction

Predict Cell Types has been extended with functionality to restrict cell type prediction to selected tissues. This improves accuracy by reducing the range of possible cell types.
Predict Cell Types matches features with versioned ids in the input (e.g. “ENSG00000243485.5”) to features with no version in the trained classifier (e.g. “ENSG00000243485”) and vice versa. Previously no match would be found for such features.
The table view of a Cell Type Classifier has been extended to contain information about the features most useful for performing the classification.

Expression plots

The heat map produced by Create Expression Plot includes ‘metadata layers’ for metadata (e.g. sample name, cluster, and cell annotations) that differs for cells in a group.
When creating expression plots, it is possible to restrict the plot to groups of cells.
When creating expression plots by right-clicking on a UMAP or tSNE plot, highlighted clusters will be pre-selected.

General

An option to set the sample name has been added to Annotate Reads with Cell and UMI. This can be set either to some literal value or a pattern in much the same way as exporters and output elements.
Annotate Reads with Cell and UMI has been extended to include two new library preparation types:
- 10x Chromium Single Cell V(D)J
- BD Rhapsody
An option to “Count intronic reads” has been added to Single Cell RNA-Seq Analysis. This supports analysis of data where many unprocessed transcripts are expected to be present, such as snRNA-seq data.
The results of Differential Expression for Single Cell additionally contain the feature id and how many cells express each feature.
New options in Differential Expression for Single Cell allow features to be filtered away when they are insufficiently expressed, prior to testing.
In the track view of an Expression Matrix, side panel options have been added to show, hide, and color features based on their expression.
The Expression Analysis from Reads workflow outputs the raw, unfiltered expression matrix. This can be used as input to the Expression Analysis from Matrix workflow, so that changes to cell calling or doublet removal can be made without having to re-map the reads.

Bug fixes

Fixed an issue where Import Cell Annotations would fail when one of the imported columns was completely empty.
Fixed an issue where the Import Expression Matrix tools could fail to import data in certain circumstances when two features had the same name or identifier. Such features are now supported provided that the combination of name, identifier, and type of feature (e.g. ‘Gene Expression’) is unique.
Fixed an issue where Import Expression Matrix in CSV/TXT Format gave an error unless the first column header was empty.
Fixed an issue in QC for Single Cell where cell annotations for “spike-in reads (%)”, “Empty droplet p-value”, and “Empty droplet FDR-corrected p-value” were missing for some cells when the tool was run on a matrix containing multiple samples. This did not affect the output filtered matrix.
Fixed an issue in QC for Single Cell where an error could sometimes be encountered when generating the histograms for the report.
Fixed an issue where tooltips shown in the UMAP/tSNE plot did not disappear in some cases after the cursor was moved.
Fixed an issue where interaction with data via the UMAP/tSNE plot would be slow when the data was generated on a QIAGEN CLC Genomics Server, and then copied and viewed locally while the workbench was still connected to the server.
Fixed an issue where p-values reported by Differential Expression for Single Cell on data normalized by Normalize Single Cell Data did not use the ‘Pearson residual’ values if a gene had 0 expression. This typically reduced power, leading to fewer differentially expressed genes being detected at a given p-value threshold.
Fixed an issue when running Create Expression Plot and Differential Expression for Single Cell from the UMAP/tSNE plot, where the pre-selected options would sometimes not reflect the selections in the side panel. This would happen if the tool was run twice, and on the first run different options than the pre-selected ones were used.
Fixed an issue where Create Subset would output clusters that did not allow easy lookup of cells in the QIAGEN Cell Ontology.

Changes

The ready-to-use workflow Perform Single Cell Analysis from Expression Matrix has been renamed to Expression Analysis from Matrix.
The ready-to-use workflow Perform Single Cell Analysis from Reads has been renamed to Expression Analysis from Reads.
Import Cell Annotations always shows an error if two or more rows in the file being imported describe the same cell. Previously an error would only be shown if the rows differed in a column containing text or integers, and in these cases the values would be imported from the last row.

CLC Single Cell Analysis Module 21.0.1

Released on March 3, 2021

Improvements

Fold changes reported by the Differential Expression for Single Cell tool on data normalized by the Normalize Single Cell Data tool were previously calculated using the normalized Pearson residual values. This could give misleading results in some cases – typically exaggerating the effect size. Fold changes on such data are now calculated by using the normalization to re-scale expressions to a fixed sequencing depth.
The cell annotations output of Predict Cell Types has been updated to only include the probabilities of a relevant subset of all possible cell types. Relevant cell types are those that are either the most likely type for at least one cell, or that have a probability above a certain threshold for at least one cell.
Predict Cell Types no longer creates empty cell types in the “Cell type (high confidence)” category of the output.
The Perform Single Cell Analysis from Expression Matrix and Perform Single Cell Analysis from Reads workflows have been updated to include the Convert Metadata to Cell Annotations tool. When a metadata table is used for configuring workflow execution, this change means that the produced cell annotations now include information from the metadata table.
In the UMAP and tSNE plot editors it is now possible to select cells that do not express a certain gene.
QC for Single Cell retains cells with metrics that are equal to the minimum values for count-based filters and maximum values for extra-chromosomal filters. These were previously considered outliers and were removed.
Various minor improvements.

Bug fixes

Fixed an issue where running “Identify marker genes” from Differential Expression for Single Cell would incorrectly identify a gene as a marker for multiple groups if the gene was a marker for one group and unexpressed in at least one other group. Affected genes appeared as markers for multiple groups with fold change +Infinity.
Fixed an issue where p-values reported by Differential Expression for Single Cell on data normalized by the Normalize Single Cell Data tool ignored negative Pearson residual values. This typically reduced power, leading to fewer differentially expressed genes being detected at a given p-value threshold.
Fixed an issue where coordinates in the table view of 2D UMAP and tSNE plots did not match those shown in the plot.
Fixed an issue where, when picking a gene in the “Expression data” side panel for UMAP/tSNE plots, the coloring would not reflect the coloring of the gene expressions until “Show colors from” in the “Coloring and highlighting” menu was changed.
Fixed an issue where running Create Subset from the right-click menu of a dimensionality reduction plot would assign incorrect normalized gene expressions to the resulting expression matrix. This affected the results produced by tools that use the normalized expressions: Cluster Single Cell Data, UMAP for Single Cell, tSNE for Single Cell, Differential Expression for Single Cell, and Create Expression Plot.

CLC Single Cell Analysis Module 21.0

Released on January 12th, 2021

First release of this module.