Question: Can I use Cell Ranger aggr to combine data from HT kits with data from standard kits? Will there be chemistry batch effects?
Answer: The Cell Ranger aggr pipeline supports the aggregation of Single Cell 3’ HT v3.1 data with data from the standard Single Cell 3’ v3.1 kit. The aggr pipeline also supports the aggregation of Single Cell 5’ HT v2 data with data from the standard Single Cell 5’ v2 kit.
With well-matched input material, we observed no noticeable batch effects during internal testing when aggregating 3’ HT v3.1 data with data from the standard 3’ v3.1 kit, or when aggregating 5’ HT v2 data with data from the standard 5’ v2 kit. For the two combinations mentioned above, we recommend starting with Cell Ranger aggr without chemistry batch correction. If the results reveal obvious differences between the HT vs. standard kit data, please check the experimental procedure carefully to identify the potential cause of batch effects. You may try Cell Ranger aggr with chemistry batch correction, which may improve the mixing of the batches in the t-SNE visualization and clustering results. However, because batch effects in these two scenarios are not expected, please use caution when interpreting results after batch correction.
If you are aggregating 3’ HT v3.1 samples with other chemistries (3’v2, 5’v2, etc.), you can treat 3’ HT v3.1 similar to standard 3’v3 chemistry. Similarly, if you are aggregating 5’ HT v2 samples with other chemistries (3’v3, 5’v1, etc.), you can treat 5’ HT v2 similar to standard 5’v2 chemistry. For example, we recommend using chemistry batch correction when aggregating 3’v3 (standard or HT) with 3’v2 chemistry. More information can be found in this knowledge base article: Can I aggregate gene expression data from different chemistries?
If you have questions about the validated usage of the chemistry batch correction in the Cell Ranger aggr pipeline, please check this page or contact support@10xgenomics.com.
Beyond the scope of 10x tools, there are a number of packages in R, such as Harmony, LIGER, scran, which attempt to address various types of batch effects. Please see this Analysis Guides article for more detail: Introduction: Batch effect correction.