Question: I would like to run cellranger vdj
and my reference is incomplete. How to use the --denovo
mode?
Answer: Denovo mode (i.e.--denovo
) is the most sought-after V(D)J assembly parameter when using a custom reference and also when there is no complete reference available (i.e. a reference with poorly annotated V(D)J genes). There are two ways you can use this parameter in cellranger commandline, and it is described below.
Method 1: using denovo mode without reference
Turning on the --denovo
argument will turn off the reference specific stages (Eg read filtering and V(D)J annotation). When cellranger vdj
is run-in --denovo
without a --reference
, the contigs are assembled denovo and it will not generate productive contigs since V(D)J gene annotations for calling productivity are not generated. Cell Ranger only considers productive contigs for clonotype grouping. Also, the .vloupe
file for importing data to Loupe V(D)J browser is not generated in this case.
- If the experiment has custom primers and if there is no reference specified, then
--inner-enrichment-primers
is required to be used along with--denovo
. - If you compare a pure
--denovo
mode without a reference and a pure--reference
mode (i.e. without--denovo
), then there will be changes in the manner in which the steps assembly, clonotype grouping, etc., are carried out, and these differences are expected. - Output files generated from a
cellranger vdj
run using denovo mode without reference with primers is seen below:
Note: Currently, in Cloud platform we do not support running Cell Ranger V(D)J in de novo mode without a reference. We are planning to add support for this option in Cloud in a future release. De novo mode without a reference is enabled in cellranger vdj/multi
command line.
Method 2: using De novo mode with reference
If --denovo
is run with --reference
, then the annotation step will use the reference but the assembly will be denovo. For instance, in a reference-assisted assembly, the annotation step in the V(D)J algorithm requires the pipeline to match the germline genes in the reference, as shown below. So for a denovo mode, this step will be performed only in the presence of a reference.
- If the experiment has custom primers, then
--inner-enrichment-primers
parameter is used along with the reference and, Cell Ranger V(D)J pipeline checks if the primers align to any genes in the reference. However, if the primers are not there, it does not get used in the assembly process of--denovo
mode and have no effect on the assembly results. - The pairing and chain information is output when using
--reference
in denovo mode. - Output files generated from a
cellranger vdj
run using denovo mode with reference with primers are seen below:
Products: Single Cell Immune Profiling