five groups with four biological replicates

[kim881030]

Hi folks.
I just began to learn how to analyze scRNA-seq data, so really basic questions might bother you.

Here is the brief scheme of my experiment.
I have five experimental groups and one group has four biological replicates. I have four hashtag antibodies and each biological replicate was hashtagged with one of four different antibodies. Four biological replicates were pooled together, sorted, and subject to scRNA-seq. I used multiplexing. After making fastq files, I ran cellranger multi to make and separate each hashtagged sample among one group. Thus, it gave me 20 final outputs.

What's the best way to find differentially expressed genes that define each group? should I aggregate the outputs from four biological replicates in each group before creating Seurat objects? or should I load and create 20 objects (or 5 aggregated files) and then merge (using merge vignette)? I want to have a big picture before doing an analysis. I hope to find a specific gene in immune cells that explain and define the target disease.

Thank you.