Yes, you can integrate Parse datasets with datasets previously generated using droplet based methods.
Through head-to-head comparisons, Parse and droplet based data have often been found to detect similar cell types and cluster identities. In light of the concordance of data, it is worth noting that higher gene detection from Parse Biosciences and lower ambient RNA can lead to high resolution of cell types. Please see the following application note using mouse lymph node nuclei to see how Parse and droplet based data integrate: Comparison of EvercodeTM WT v2 and ChromiumTM Next GEM Single Cell 3’ Kit v3.1 in Mouse Lymph Node Nuclei.
Additional comparisons, including for mouse brain nuclei samples, are available on our website at: https://www.parsebiosciences.com/datasets/
Note that to best merge datasets, data integration approaches may be useful. Similar to when profiling samples from different biological donors or when profiling samples that were processed on different droplet-based scRNA-seq runs, there can be technical and non-biological variation. Integration is an approach that makes it possible to match cell populations across datasets where batch effects may exist to conserve biological variation.
You may refer to this benchmarking paper from the Theis lab (2022) to explore the different tools available and pick one that best suits your needs (size of dataset, computational resources available, characteristics of datasets - common or different cell types). Examples of some of the tools discussed in this paper are Harmony, BBKNN and Seurat v3. You can find a tutorial here that uses Harmony to remove batch effects in a dataset with 24 samples.
There are a few considerations to keep in mind while integrating existing datasets with Parse datasets:
1. Downsample the number of reads per cell per sample so the read depth from both platforms are comparable. This will require you to re-run the pipeline with the downsampled sublibraries. If you have samples with varying read depths in the same experiment please get in touch with support@parsebiosciences.com for additional information on how to best handle these datasets.
2. If your Parse assay included multiple samples, the 'all-well' or 'all-sample' count matrix will encompass all these samples. Droplet based methods usually generate gene count matrices for each individual sample. In order to make sure you are working with the appropriate sample, please start with the sample-specific count matrix.
SL1-out/sample_1/DGE_filtered/
SL1-out/sample_1/DGE_unfiltered/
3. Ensure that the same reference genome build has been used for analysis of both experiments.
Please note that if you have samples with a variable number of cells, this can affect the number of sub-populations of cells you see in each sample, in spite of integration.