Different RNA-Seq experiment types require different sequencing read
lengths and depth (number of reads per sample). This bulletin reviews
RNA sequencing considerations and offers resources for planning
What resources should I consult first?
For RNA sequencing, read depth is typically used instead of
coverage. Detecting low-expression genes can require an increase in
read depth. The ENCODE
project has data standards for RNA-Seq
RNA sequencing that are an excellent resource for many projects.
Illumina recommends consulting the primary literature for your field
and organism for the most up-to-date guidance on experiment design.
How many reads should I target per sample?
Read depth varies depending on the goals of the RNA-Seq study. Most
experiments require 5–200 million reads per sample, depending on
organism complexity and size.
- Gene expression profiling experiments that are looking for a
quick snapshot of highly expressed genes may only need 5–25 million
reads per sample. In these cases, consider pooling multiple RNA-Seq
samples into one lane of a sequencing run. This allows for high
multiplexing of samples.
- Experiments looking for a more
global view of gene expression, and some information on alternative
splicing, typically require 30–60 million reads per sample. This
range encompasses most published RNA-Seq experiments for mRNA/whole
- Experiments looking to get an
in-depth view of the transcriptome, or to assemble new transcripts,
may require 100–200 million reads. In these cases, consider
sequencing a few samples across several high output sequencing
- Targeted RNA expression requires fewer reads. For
example, Illumina recommends 3 million reads per sample for TruSight
RNA Pan Cancer and TruSight
RNA Fusion Panel, which are compatible with high plexity
pooling of samples.
- miRNA-Seq or Small RNA Analysis
experiments may require even fewer reads than whole transcriptome
sequencing. This requirement varies significantly depending on the
tissue type being sequenced. Illumina strongly recommends using the
primary literature to determine how many reads are needed, with most
applications ranging from 1–5 million reads per sample.
To determine how many samples can be run at one time, divide the
number of reads produced by the flow cell, by the number of reads
needed per sample.
number of reads per flow cell / number of reads per
sample=number of samples per flow cell
How long should my reads be?
Read length will depend on the application and final size of the
library. The Library
Prep Kit Selector provides read length guidance for each type of
RNA-Seq library. Sequencing reads that are longer than the insert
length do not provide additional useful data.
- Gene expression / RNA Profiling – Quantifying the coding
transcriptome typically requires a short single read (often 50–75
bp) to minimize reading across splice junctions while counting all
RNAs in the pool.
- Transcriptome Analysis – Novel
transcriptome assembly and annotation projects tend to benefit from
longer, paired-end reads (such as 2 x 75 bp) to enable more complete
coverage of the transcripts and identification of novel variants or
splice sites. Paired-end reads are required to get information from
both 5’ and 3’ ends of RNA species with Stranded RNA-Seq library
- Small RNA Analysis – Due to the short
length of small RNA, a single read (usually a 50 bp read) usually
covers the entire sequence. A read length of 50 bp sequences most
small RNAs, plus enough of the adapter to be accurately identified
and trimmed during data analysis.
For more information on other considerations in planning your
RNA-Seq experiments, and RNA-Seq kit options, see our recorded
RNA Sequencing Part I: Introduction
Part II: Kits
III: Best Practices
Part IV: Introduction to Analysis
Small RNA-Seq Part I: Introduction and Part II: Best Practices
For read depth considerations per instrument, refer to the Illumina