Question: Why are the "Reads Mapped Confidently to the Probe Set" or "Reads Mapped Confidently to the Filtered Probe Set" metrics in my Spatial Gene Expression web summary file low?
Answer: In a typical Visium Spatial Gene Expression for FFPE web summary file, we expect to see the “Reads Mapped Confidently to the Probe Set” and "Reads Mapped Confidently to the Filtered Probe Set" metrics to be >50%. Lower values of these metric are unexpected and are typically associated with the presence of single probes, mismatched probe pairs, or off-target probes in the final library. Outlined below are the common root causes of these metrics being low and mitigation strategies to maximize the likelihood of success with the Visium for FFPE workflow.
Low Fraction Reads Mapped Confidently to the Probe Set:
Observing a “Low Fraction Reads Mapped Confidently to the Probe Set” flag in the web summary file generally suggests that there was a workflow error; such as suboptimal Post-Hybridization/Ligation Washes (e.g. deviation in the temperature, timing, or the number of washes) or tissue detachment. Lower mapping rates can be further compounded by lower RNA quality. Depending on the fraction of usable reads and complexity of the tissue of interest, viable data may still be generated with lower “Fraction Reads Mapped Confidently to the Probe Set” values.
Low Fraction Reads Mapped Confidently to the Filtered Probe Set:
The Probe Set CSV reference consists of probes that are either included in the analysis (designated as “TRUE”) or excluded (“FALSE”) by default. Probes with the “FALSE” label have predicted off-target activity. Probes with the “DEPRECATED” prefix are always excluded from the analysis. Observing a “Low Fraction Reads Mapped Confidently to the Filtered Probe Set” flag in the web summary file suggests that your block of interest expresses transcripts that off-target probes (labeled as “FALSE”) detect. Depending on the fraction of usable reads and complexity of the tissue of interest, viable data may still be generated with lower “Fraction Reads Mapped Confidently to the Filtered Probe Set” values. Values of the “Reads Mapped Confidently to the Filtered Probe Set” metric are normally specific to the tissue type and/or block.
See these Q&A articles for additional information on the probe sets:
- Is my gene in the Visium for FFPE probe set?
- How can I change spaceranger filtering for FFPE probes?
Additional considerations:
We recommend using DV200% to assess RNA fragmentation before proceeding with Visium for FFPE because through in-house experiments, we’ve found that the DV200% metric and assay sensitivity generally appear to be correlated. However, DV200% is not a perfect predictive tool of assay performance. In some cases, we’ve seen tissue blocks with high DV200% have suboptimal performance. Additional properties of block quality beyond RNA fragmentation may contribute to performance, such as complexity of the tissue, its compatibility with our current decrosslinking/permeabilization conditions, or adhesion.
Products: Visium for FFPE; Visium CytAssist for FFPE