FAQ

1. Why is my reported mtRNA % zero?

DRAGEN StrataMap calculates mitochondrial RNA (mtRNA) by matching chromosome names in the genome's GTF file against a list of expected mitochondrial chromosome identifiers. By default, the pipeline looks for chromosomes named chrm, m, or mt (case-insensitive).

If your mtRNA metrics show zero or unexpectedly low values, the most likely cause is that your genome uses a different naming convention for the mitochondrial chromosome.

The pipeline reports which genes it found in the GTF and used to calculate mtRNA in the following file:

intermediate_results/04_matrix_metrics/[samplename]/[genome]_mtRNA_genes.csv

How to fix this:

You can customize which chromosome names the pipeline should recognize as mitochondrial by adding the mitochondrial_chr parameter to your analysis config file:

1. Create or edit your analysis_config.json file.
2. Add the mitochondrial_chr parameter with a comma-separated list of chromosome names that should be considered mitochondrial in your genome:

{
    "mitochondrial_chr": "NC_041761"
}

1. Pass this config file to the analysis using the ICA UI.

2. Why is my mtRNA lower than expected for a custom genome?

Even when the pipeline correctly identifies the mitochondrial chromosome, we sometimes observe a low percentage of mitochondrial counts reported for custom genomes. This is due to NUMTs (nuclear mitochondrial DNA segments), which are segments of mitochondrial DNA inserted into a nuclear chromosome. NUMTs arise when mitochondrial DNA escapes the mitochondrion, enters the nucleus, and integrates into a chromosome.

They are common across vertebrate genomes. Because NUMTs are derived from the mitochondrial genome, they share high sequence identity with it, causing RNA-seq reads from genuine mitochondrial transcripts to map equally well to both the mitochondrial chromosome and the nuclear NUMT copy. Short-read aligners will mark these as multi-mappers, and if they are discarded, the result is an artificially low mitochondrial gene count.

For example, the Rhesus macaque mmul10 genome has the entire mitochondrial chromosome copied into chromosome 8.

In version X of the pipeline, the ability to include multimapping reads was added. Because this prioritizes protein-coding alignments over pseudogenes, altering the annotations of the mitochondrial chromosome genes relative to the NUMTs could help solve this issue.

3. Why are so many of the cells in the NA cluster?

By default, DRAGEN StrataMap filters out cells with >40% mtRNA. Depending on the sample type, this may remove a large fraction of cells. For example, in mouse heart samples, more than 40% of cells may have >40% mtRNA. To adjust the default threshold:

1. Create or edit your analysis_config.json file.
2. Add the mt_threshold parameter with a higher percentage:

{
    "mt_threshold": "50"
}

1. Pass this config file to the analysis using the ICA UI.

4. Why is the % of reads mapping to the genome so low?

Several factors can reduce the percentage of reads that map to the genome:

• Custom genomes may be incomplete, such as missing the mitochondrial chromosome, or poorly annotated. This can reduce the number of reads that align or assign to genes.
• Poor sample quality and/or over-permeabilization can produce degraded RNA. These fragments may be too short to align successfully. This is often associated with high trimming rates and a large number of reads being discarded during trimming.
• Alignment parameters choices such as excluding multi-mapping reads and intronic reads can decrease mapping rates. The size of effect of these settings have depends on the species, reference genome and tissue type.