What is the recommended sequencing depth for SLAMseq-QuantSeq experiments?

The ideal sequencing depth for your SLAMseq experiment needs to be adjusted to the S4U incorporation efficiency, which is specific to your experimental conditions (e.g., S4U labeling time, S4U incorporation rate, and cell type). To determine the required sequencing depth, we recommend a 2-step sequencing approach:

1. For a first test screening, sequence a SLAMseq-QuantSeq library mix as e.g., part of a lane mix. This spike-in run aimed at low sequencing depth (1 - 3 million reads) should enable you to detect some T>C mutations in these reads. Based on these results, you can look at the percentage of the total reads coming from these libraries to judge what level of coverage is required per sample.

2. Sequence the SLAMseq-QuantSeq libraries to the required sequencing depth by pooling reads from test screening and completion run. Therefore, subtract the extra sequencing depth, which was obtained from the first spike in run, from the target read depth for the stand-alone sequencing run(s). This staged sequencing approach minimizes the risk to sequence at too high sequencing depth.
   

Follow-up experiments can be sequenced directly at the desired sequencing depth. Regarding sequencing format, we recommend single-read 100 -150 bp (SR100-150) sequencing runs.

As a minimum read depth, we recommend at least 20M reads per sample as a starting point. However, the ideal depth might depend on the timepoints used. For example: For very early timepoints the overall percentage of newly synthesized RNA is much smaller than at later timepoints, hence greater depth might be needed if you are working on shorter time scale vs longer time scales. Therefore, a spike-in run (as described above) is highly recommended to judge the required sequencing depth.

The developers of the SLAMseq method recommend the following in their published research: As a rough estimate (based on experiments using mouse embryonic stem cells), we recommend for longer labeling times (i.e. >3h) 10-20 million and for short pulse labelling (i.e. <3h) 30-50 million reads per library for efficient quantification of s4Ulabeled transcripts. Herzog et al., "Thiol-linked alkylation of RNA to assess expression dynamics" in Nature Methods, DOI: 10.1038/nmeth.4435