Cell Ranger v8.0 (and later) requires users to specify the --create-bam parameter while running the cellranger count and cellranger multi pipelines. The --create-bam parameter replaces --no-bam of Cell Ranger v7.2 and earlier.
Usage
--create-bam=<true|false>: Enable or disable BAM file generation
For the multi pipeline, you can specify create-bam,true or create-bam,false in the multi config CSV as follows:
[gene-expression]
reference,/path/to/transcriptome
create-bam,true
Setting parameter to FALSE:
--create-bam=false disables the generation of BAM files.
Setting the option to false reduces the overall computation time and decreases the size of the output directory, as it omits the generation of the BAM file. The decision to skip BAM file generation has pros and cons, and these considerations vary depending on the type of assay.
In Fixed RNA Profiling (Flex), gene counts are determined by counting barcodes linked to a probe set. Unlike traditional methods, Flex does not require the alignment of reads to the transcriptome for counting. Opting to generate a BAM file introduces an additional step, wherein reads from the probe insert are aligned to the transcriptome using STAR. This extra step contributes to an increase in computation time. Given that these are probe sequences, the information derived from the loci and their associated sequences is entirely redundant with probe counts, incurring a significant computational overhead.
For 3’ and 5’ Gene Expression assays, aligning reads is a pivotal step for gene-associated read counting. Unlike Flex-based assays, this alignment step is integral to the pipeline. The primary advantage of forgoing BAM creation in these assays is a reduction in output size, although there is also some decrease in compute time. This trade-off is particularly beneficial when a smaller output size is prioritized over the detailed information encapsulated in a BAM file.
Setting parameter to TRUE
--create-bam=true enables the generation of BAM files.
Creating BAM files can be advantageous in specific scenarios. For instance:
- Third-party analysis tools: Some third-party analysis tools may necessitate BAM files as input (e.g., variant calling, RNA velocity analysis).
- Regenerating FASTQ data: If there is a need to regenerate FASTQ sequencing data using the bamtofastq pipeline, having BAM files is essential.
- Troubleshooting: BAM files can serve as valuable assets for troubleshooting purposes in certain instances.
To err on the side of caution, especially if you are uncertain about your specific needs, it is recommended to set --create-bam=true. This ensures that BAM files are generated. It is also worth noting that the 10x support team may request BAM files to assist you in troubleshooting certain issues. Therefore, having these files readily available can streamline the support process.