Handling Multiple Sequencing Runs in scRNA-seq

Question

[dk0319]

I am interested in analyzing some scRNA-seq data from GEO and I found that the samples have multiple runs (4 in some cases). I am wondering if someone has experience with this and if there is a recommended standard procedure for handling multiple runs?

For ChIP or bulk-rna seq I would just merge the files but I am unsure if this is normal for single-cell data.

[tothepoint]

Please check if the four sample named in pattern L001/001/003/004 which can be understood by sequencing was done in 4 lane of the sequencer and you can merge the fastq file to run your downstream analysis.

[dk0319]

This is how the metadata is presented on SRA explorer. Just to confirm I should be able to merge these as 1?

"SRX15246019: GSM4274678: BMET1-Tumor; Homo sapiens; RNA-Seq 4 ILLUMINA (Illumina HiSeq 4000) runs: 130.2M spots, 12G bases, 5.3Gb downloads"

Runs: 4 runs, 130.2M spots, 12G bases, 5.3Gb"
Run # of Spots  # of Bases  Size    Published
SRR19180790 32,510,181  3G  1.3Gb   2022-05-17
SRR19180791 31,813,145  2.9G    1.3Gb   2022-05-17
SRR19180792 32,963,341  3G  1.3Gb   2022-05-17
SRR19180793 32,918,198  3G  1.3Gb   2022-05-17

[rpolicastro]

It says it's RNA-seq.

[dk0319]

Yes, but the data is scRNA-seq with R1, R2 and L1 fastq files

[dk0319]

You are correct this is how the runs are labeled as L001-004. My question now is because each of these samples consists of R1, R2 and L1 fastq files, am I correct in assuming I need to merge all the R1s and R2s and L1s individually to produce 3 final fastq files?

[Marco Pannone]

Do you mean the samples come from different batches? In that case, you want to account for batch effect. You can find a nice explanation on how to account for batch effect here.