Single Cell Analysis Workflows

March 28, 2023 20:56
Updated

Single Cell or Single Clone Analysis workflows are useful for anyone who is expecting to find one (or more) dominant heavy and light chains in each of their samples. This could be barcoded data where each barcode corresponds to a dominant heavy and light chain pair, or if you have fasta/fastq files where each file is expected to have a dominant chain or pair of chains. This covers a range of different types of experimental processes, including both Single Cell and Single Clone workflows. See Understanding Single Cell technologies: Barcodes and UMIs if you are unsure what single cell analysis is.

There are three main ways to handle 10X and other barcoded/UMI sequence data in Geneious Biologics. The best path depends on how much preprocessing has already been done prior to upload, which will determine your entry point.

Method 1 - Pair, Collapse UMIs, Single Clone Analysis

If you have a fastq/fasta (or other format) file with raw sequences that contain UMI and/or Barcodes and have not been demultiplexed, this is the method to use. The NGS Tutorials 3 & 4 on our website follow a similar workflow, so that may be a good place to start. You can find them here: NGS Tutorials 3: Using Barcodes and UMIs and Tutorial 4: Single Clone Antibody Analysis. This tutorial has a small sampled dataset, the full dataset can be provided on request.

A typical workflow could look like:

Set + Merge Paired Reads: feed raw FASTQ files into 'Set and Merge Paired Reads'. You may wish to choose the "Set pairs only" option if your merge rate is low. The UMI/Barcode tool can handle paired unmerged reads.
UMI Collapse and Separate Barcodes: Feed paired (or merged) reads into 'Collapse UMI duplicates and Separate Barcodes'. This merges all similar sequences with the same UMI and Barcode together. It is useful to leave the Trim UMIs/Barcodes' option off while experimenting, but you will want to have them trimmed off for the next step. The Barcode will be saved on each sequence as metadata. See Collapse UMI Duplicates and Separate Barcodes.
Annotation: Feed the (trimmed) generated consensus sequences into 'Single Clone Antibody Analysis'. Single Clone Annotator is a variation of Antibody Annotator that attempts to combine related sequences within a barcode so that you can see the dominant clones. You could also use Antibody Annotator if you wish, which will analyse each of the sequences without merging them. Both tools offer clonotype analysis (our "cluster" tables, see Understanding "Clusters").
Discovery: Alignments can be done directly from the results so that the aligned sequences are annotated. You can choose to align translated regions of interest, with trees.
Export: Extract selected annotate candidate sequences, or export tables, graphs, and images for reporting or record keeping purposes. See Exporting Annotated Sequences and Sequence Tables.

Method 2 - Single Clone Analysis from Separate Lists

If you or your sequence provider have already done some demultiplexing in another tool, you can feed the preprocessed sequences into Geneious Biologics for annotation and analysis. If you have your sequences sorted into sequence lists (such as one fastq file per clone/barcode/well/cell) then you could use these sequences directly with Single Clone Analysis. When you run Single Clone Analysis on multiple files or sequence lists, then it will attempt to find a dominant heavy and light chain for each list. NGS Tutorial 5. Single Clone Analysis from Separate Lists goes over this method.

Method 3 - Antibody Annotator + Add Assay Data

If you have already done some demultiplexing in another tool such as Cell Ranger, but don't have them sorted into sequence lists, then you could feed the consensus sequences directly into Antibody Annotator. You can import it as sequences (such as fasta, fastq or genbank format) or as a csv file using our csv sequence import tool.
Antibody Annotator will annotate FR and CDR regions, identify liabilities, variants relative to germline, clonotypes and more. The "Add Assay Data" feature can also be useful to add any coverage or depth information from csv or excel files from your demultiplexing stage.

Note: For barcoded data in particular, when trialling it can be useful to have actual samples of your expected data format. This allows you to determine the optimal pre-processing and analysis workflow and make tweaks where necessary. As always, please contact us if you have any questions or feedback.

FAQ: Why do my results for Antibody Annotator and Single Clone Analysis look different?

Single Clone Analysis and Antibody Annotator have some differences in their underlying algorithm, so it is not surprising if their results are similar but not exactly the same. The inherent differences in the analysis are explained below, so that you know how to compare the two results.

Antibody Annotator

Antibody Annotator annotates sequences individually or in pairs (for Illumina paired reads). That means each individual input sequence gets its own row in the All Sequences table, even if it is only a fragment of a variable region or not identifiable. The cluster tables shows the number of sequences that have the exact same protein sequence over the chosen region. For the VJ cluster table, this will only show sequences (or chain pairs!) that were identified as having a full VJ region. So the VJ Cluster table shows the number of sequences with the exact same protein sequence over a complete VJ region. Read the Understanding "clusters" article for more details.

Single Clone Analysis

Single Clone Analysis is designed to find dominant sequences within a sample or samples, and is often used when the scientist expects to only find one or two dominant heavy/light chains per sample/cell/clone/sequence list/barcode. It is often used in combination with barcoded sequences, or for batch analysis where each uploaded file represents a different but related sample. Single Clone uses some special rules:

A single row in the "All Sequences" table may represent many individual reads.
Sequences are combined together based on a similarity threshold (you can set this in the options). If you have the 'De Novo Assembly' option turned on then sequences are assembled together. This means that one row in the All Sequences table might include sequences that originally had slightly different VJ protein sequences. The VJ Region amino acid sequence shown is the consensus across all similar reads.
Only "Fully Annotated" sequences (or assembled contigs) are kept to show in the analysis results. You can define what regions should be used to determine "Fully Annotated".
Sequences shorter than a certain length may also be filtered out, depending on your options.
More detailed explanation of Single Clone Analysis filtering and options can be found in the main Single Clone Article.

Summary: There are a number of differences between the two analysis pipelines, but the main difference is that for Antibody Annotator the VJ Regions you highlight above are showing sequences with the exact same VJ protein sequence, whereas with Single Clone it is showing a consensus of similar VJ regions - a reduced dataset.