Gene causes optical crowding in Xenium panel design

Question: A gene is causing optical crowding in my Xenium panel design. What should I do?

Answer: First, try reducing the probe set coverage for the gene, e.g. to four and then to two if needed. In the Cloud Panel Designer, explicitly specify the reduced probe set coverage in the same row where you input the gene by separating the coverage with a comma, e.g. like so:

Several rounds of iteration are likely necessary to maximize the probe set coverages while keeping the per-gene and per-cell-type TP10K utilizations within the acceptable thresholds.

• In general, the panel designer aims for eight probe sets per add-on gene if possible. Some genes may enable fewer probe sets. 
• In general, having three probes for a gene is safe, but having two or one probe for a gene runs the risk of not being detected. Premade base panels include genes with one or two probe set coverages.
• If a gene with one probe is still showing overly high utilization then we recommend removing it. Substitute with a covarying gene with lower expression. 

If the gene is essential to research, and if the reduction for the gene is insufficient to reduce the predicted optical crowding, then the next approach to try is global coverage reductions that reduce the probe set coverages of the other genes for the cell types predicted to have optical crowding. For example, take the next 10-20 highly expressed genes in such a cell type and reduce their probe set coverages. Alternatively or additionally remove the bottom low expressing genes. The incremental reductions can add up to offset the utilization by the high expressing gene.

Per-gene, the limit is 120 TP10K utilization for any cell type. The per-cell-type limits differ for Xenium chemistries as follows.

Having a handful of cell types with TP10K utilization between the conservative and absolute cutoffs is acceptable. This is more acceptable for rarer cell types than more abundant cell types. Widespread coverage reductions will reduce sensitivity in detection for the genes. This should be weighed against the sensitivity loss due to optical crowding.

• It is more acceptable for rare cell types to go above the conservative threshold.
• The conservative threshold is where the loss due to crowding is minimal.
• Reducing coverage for so many genes will have an impact on observed sensitivity.
• Above the conservative threshold optical crowding effects will dominate.

An exception to the above is base panel genes that cannot be removed and are highly specifically expressed in a limited space, e.g. in rare cell types. An example is insulin-expressing islet cells in the  pancreas. The human multi tissue base panel includes INS, and this gene will show extreme optical utilization for pancreatic islet cells. In such a case, it is okay to continue with the design.

Related resources

1. Xenium Panel Designer documentation homepage
2. Interpret Xenium Panel Design Summary
3. Xenium Panel Designer Frequently Asked Questions
4. Pre-designed Xenium V1 Gene Expression Panels
5. Pre-designed Xenium Prime Gene Expression Panels
6. Xenium Add-on Panel Design Technical Note
7. Species Standalone Custom and Advanced Custom Panel Design for Xenium In Situ Technical Note
8. Creating Single Cell References for Xenium Custom Panel Design from Seurat or AnnData
9. Xenium Panel Designer Examples (10x Cloud)

Products: In situ Gene Expression

Last updated: 8/29/2024