Question - How can I demultiplex Visium libraries with low quality index reads?
Answer - After sequencing your libraries, in some instances, you may be faced with low quality bases (or long Poly G stretch) in the indices of your Visium samples. This could be either the Index 1 (i7) or Index 2 (i5) reads. To generate FASTQ files, the demultiplexing considerations are discussed below for libraries that display this phenomenon.
- As the libraries are dual indexed, there may be a desire to demultiplex based on just one index sequence. This is not advised for Visium samples due to the low barcode complexity (5k - 14k spot barcodes).
- If there are multiple libraries pooled on a lane, there is a potential for index hopping and the samples may end up mixed or contaminated. Single index demultiplexing does not account for index hopping. When multiple libraries are run from different tissues on the same flow cell lane with a single index, we have observed cell-type specific markers from both tissues in both samples. Single index demultiplexing is not recommended for Visium libraries and is not officially supported by 10x Genomics.
- The index hopping filter software tool is not suitable for Visium libraries if the samples were pooled together on the same flow cell.
- If the sample is pooled, the only available option is to use the demultiplexed reads from both Index 1 and Index 2 to ensure high quality data. There may be loss of reads to the “Undetermined” category because of the drop in quality. However, the remaining demultiplexed reads identified by the Index 1 and Index 2 sequences provided in the sample sheet are reliable and can be used for analysis.
- If the library is not pooled with others on a lane, then single index demultiplexing may be an option to recover these data.
Please contact support@10xgenomics.com for additional information.
Products: Visium FFPE v1 and Visium CytAssist FFPE assays.
Related articles
Can Visium for FFPE libraries be pooled with other dual indexed 10x libraries?
Can Visium CytAssist Spatial Gene Expression libraries be pooled with other 10x libraries?