Xenium In Situ measures gene expression in tissue sections derived from either formalin fixed & paraffin embedded (FFPE) or fresh frozen (FF) tissue samples placed on Xenium slides. Gene expression is measured with panels of probes that target genes of interest.

10x Genomics offers several customization options for Xenium In Situ Gene Expression panel designs. The pre-designed and custom panel types are described below.

• If you have decided to use pre-designed panels, please see the pre-designed panel information and part number pages for Xenium v1 and Xenium Prime 5K options.
• Custom design options are implemented by the Xenium Panel Designer tool. If you have decided to pursue a custom panel for your experiment, please read the Xenium Panel Designer Workflow Guide page for guidance on assay chemistry compatibility and tool workflow steps:

For new Xenium users, 10x Genomics recommends starting a pilot experiment with pre-designed panels. There are several advantages to this approach:

• Pre-designed panels have been extensively tested by 10x for sensitivity and specificity on the specific tissue types.
• Pre-designed panels are in inventory, so researchers can get started right away. No need to wait to design and manufacture a custom panel.
• Pre-designed panels are well-suited for exploratory analyses that identify major cell types in the selected tissue.
• Researchers can ensure their lab is comfortable with the entire workflow, and their first datasets can be compared to 10x Genomics public data demonstrations.
• It is easier to reproduce experiments across labs or research groups.
• If there is any experimental optimization or troubleshooting to be done with the 10x Genomics Support Team, starting with pre-designed panels may help by first addressing any issues unrelated to specific panels.

If you are only using 10x Genomics pre-designed panels, no need to use the Xenium Panel Designer! Learn about available pre-designed panels here:

• Xenium v1 panels
• Xenium Prime panels

10x Genomics offers several types of custom panel design.

• Researchers may be interested in selecting additional markers for specific subpopulations of cells, or genes expressed under specific conditions (e.g., disease), that are not available in pre-designed panels. You may add on to a pre-designed panel or request a standalone custom panel that is independent from pre-designed panels.
• If your experimental goals do not align with the above options, please contact your sales representative for information about advanced custom targets (i.e., specific gene isoforms) or species standalone requests (non-human/non-mouse species).

Add-on custom panel designs offer the flexibility to include more genes to available pre-designed panels (see Xenium Add-on Panel Design Technical Note). The add-on custom configuration is only compatible with pre-designed panels; it is not compatible with a standalone custom design.

Review the curated pre-designed panel gene lists carefully. The lists are designed to comprehensively cell type a sample. Do the genes adequately cover the biology you are exploring? If not, consider adding additional genes to complement it.

To test a biological hypothesis, researchers may want to select markers for certain subpopulations of cells or genes expressed under specific conditions (e.g., diseased vs. healthy). For example, 10x scientists designed an add-on panel to better understand breast cancer biology in this publication: "High resolution mapping of the tumor microenvironment using integrated single-cell, spatial and in situ analysis". They analyzed data from 313 genes total: 280 genes from common cell types using the pre-designed Xenium Human Breast Panel and 33 additional genes, like invasive myoepithelial marker genes. The add-on genes were selected and curated primarily based on single cell reference data for both healthy and tumorigenic human breast tissues (Pal et al. 2021, Bhat-Nakshatri et al. 2021, Karlsson et al. 2021) (see Supp. Fig. 1).

Standalone custom panels may be useful when the human or mouse tissue being analyzed has a very different expression signature compared to the specific tissues covered by pre-designed panels.

We support standalone custom panels up to 480 genes. However, depending on the desired gene list and tissue(s) being analyzed, the final number of genes may be reduced in our design process to ensure high sensitivity across all biology of interest. Thus, we recommend submitting ~10-20 additional genes that could be used in the design process to replace genes in the main set as needed.

Advanced targets may be included in custom design requests. Possible advanced targets include:

• Exogenous sequences: Use this category when you want to target a sequence with multiple probes (e.g., fluorescent reporters, viral transcripts, bacterial transcripts, protein tags, transgenes, and xenografts - see below). If you need to target a specific location with a larger sequence (e.g., a splice junction) use the category Junction sequences.
• Junction sequences: Use this category when you want to target a specific sequence such as a splice junction (e.g., gene fusions, isoform junctions).
• CRISPR guide RNAs
• CDR3 clonotypes
• Barcodes (e.g., lineage tracing)
• SNVs or indels
• Other: Use this category for requesting targets that are not listed. Selecting Other will initiate a conversation with our commercial and R&D team to determine if we can support your desired application.

Additional information about custom target design considerations is provided below.

Exogenous sequences

• Fluorescent reporters, transgenes, and protein tags: All types of exogenous sequences can be targeted with Xenium In Situ chemistry. An important consideration when designing against exogenous sequences is that these sequences are often highly expressed, which can cause optical crowding. A 10x scientist will work with you to reduce the risk of optical crowding.
• Bacterial sequences: Some researchers have successfully targeted both bacterial and human/mouse sequences in a sample. The biggest predicted risk is that the expression level may be very high, causing optical crowding. While designing probes for bacterial sequences is possible, sample preparation for these experiments has not been tested. RNA accessibility in bacteria can be biased by the sample preparation protocol, with some bacterial species being especially difficult to lyse.
• Viral sequences: Some researchers have successfully targeted both viral and human/mouse sequences in a sample. It is advised to design probes that target transcribed RNA. While it is technically possible to design probes that target ssRNA viruses directly, secondary structure and RNA copy number may lower sensitivity. Additionally, viral RNAs can have a wide dynamic range of expression, which could lead to either low sensitivity or optical crowding on either extreme.

Junction sequences

Isoforms and gene fusions: Isoform or gene fusion designs are similar in that they both target a specific junction of sequences. Preliminary data suggest that isoform detection with Xenium In Situ chemistry is robust. Probes can be designed to target each splice junction of interest. The Xenium Panel Designer app and 10x scientists consider optimal ligation junction sites (+/- 5 bp from splice junction). If no optimal ligation junctions exist within 5 bp, they will try other junctions. Analysis of the mouse and human transcriptomes suggest that >99.9% of splice junctions have a ligation junction within five bases of the splice junction.

CRISPR guides

CRISPR guides can be detected directly with Xenium In Situ chemistry, or indirectly through co-expressed barcodes (see below). 10x has previously demonstrated how to detect CRISPR guides directly using the Flex Gene Expression assay (see poster). The probe design considerations for the Xenium In Situ platform are similar.

CDR3 Clonotypes

T cell receptor (TCR) detection in Xenium In Situ chemistry can come in one of two general flavors: 1) detection of Complementarity-Determining Region 3 (CDR3) sequences or 2) unbiased profiling of V/J segments. CDR3 sequence detection relies on using an assembled CDR3 sequence to design probes specific for the target CDR3 (such as one produced by the 10x Immune Profiling solution). By doing this, we make the assumption that there are no other CDR3 present at appreciable levels in the sample that have close sequence similarity.

Barcodes

Xenium In Situ chemistry can detect any ssRNA at sufficient copy number. We recommend at least 2-4 copies per cell, but ideally ≥ 10 copies. The ideal number of copies per cell for detection will depend on the proportion of cells expressing your barcode and the efficiency of the probe.

A common use case for barcode detection is lineage tracing. Some use cases involve multiple barcodes in a combinatorial fashion. When considering an experimental design that relies on multiple targets being detected in a given cell, you should consider the combined probability of each probe detecting a target when present. We suggest no more than 2-3 barcodes per cell to ensure that there is a reasonable probability you detect all barcodes when they are present in a cell.

Placing multiple copies of the same barcode on your construct will increase sensitivity by boosting the effective copy number. If you want to design probes in this fashion, it is ideal to put 50 bp of space between each barcode copy.

SNVs or indels

Xenium In Situ chemistry can be used to detect variants as small as a single nucleotide. The most common design method is to design two independent probes for each haplotype, which bind competitively to boost specificity. These two probes will be on separate codewords, which means each SNV/indel uses two targets in your panel of the up to 100 add-on custom targets.

Due to the high similarity of the probes, it is expected that there will be mis-ligation roughly 10% of the time, and this should be taken into account in your downstream analysis. Additionally, the performance of SNV probes is hard to predict. Due to the design constraints where the site of the variant is placed right at the ligation junction, it may not be possible to design probes with good sensitivity and specificity at the same time. Targeting many closely spaced variants will also further reduce sensitivity.

Xenografts

Xenografts are transplants of organs, tissues, or cells to another species, most commonly human and mouse. 10x Genomics has built a set of gene expression designs that take human and mouse sequences into account. In general, 10x expects that multi-species experiments should work, but may require specialized downstream analyses to identify the species of a given cell.

In XPD, the selected organism (Human, Mouse, or Other) will determine the pre-designed panel genes of the custom design. To add genes from another organism to the pre-designed base panel, you can select "Exogenous sequences" in the Advanced target section and upload the list of genes in the XPD workflow (i.e., mouse genes on a human pre-designed panel, or vice versa). Non-human or non-mouse xenograft requests are enabled for Xenium v1 experiments by selecting "Other" as the organism type and "Other" as the Advanced Target.

Species standalone custom panels are panels that target endogenous gene expression in species beyond those currently offered by 10x Genomics, or panels that include multiple species. If you are interested in a non-diploid species, contact your Sales Executive. These options are only offered for Xenium v1 chemistry experiments.

To profile endogenous gene expression in eukaryotic species, it is required to have a high-quality genome assembly and annotations with a well-matched single cell reference. 10x Genomics does not offer supported sample preparation protocols for species besides human and mouse. 10x expects human and mouse protocols to be adaptable to corresponding tissue types for other mammals, particularly rodents and primates, but this is not guaranteed. Researchers working with increasingly divergent lineages from mammals, particularly fish, insects, and plants, should carefully consider the impact of their unsupported sample preparation protocols on overall experimental success.

Researchers will also need to consider downstream analysis needs. For example, Xenium cell segmentation algorithms have not been tested, and are not supported, on datasets beyond specific human and mouse tissues. Researchers should be prepared to segment cells manually or with third-party tools.