I performed PE100 sequencing of 150 RNA-seq samples on a NextSeq S4 flow cell. However, we did not reach the minimum required depth (minimum -70 million reads, desired - 80 million reads) for 45 of those samples. The median number of reads we are missing to reach 80 million reads for those 45 samples is 15 million (although 6 samples need 30 million reads).
I was considering whether I could run a PE100 sequencing run in a new flow cell for library samples for which I did not get the desired number of reads and concatenate the .fq files obtained from this new sequencing run with the .fq files from the previous Novaseq S4 sequencing run. I read an article by Maroni et al 2008 that said this was fine (considering they use the same technology).
These samples will be used for detection of variants in genes expressed in tumor samples. I've read that merged .bam files increase the sensitivity of SNP calling (doi: 10.1155/2014/319534), but also that it might increase the number of False positives (doi: 10.1038/s41467-019-09026-y).
By following this strategy, might information be lost in genes with low expression levels? I don't see that there could be a problem as we would be reaching the required depth in the end. Can this false positive rate be problematic? At what step should I merge technical replicates .fq or .bam files?
For further information, I would want to perform the new sequencing run in a different sequencer. The original run was on a Novaseq-6000. I would like to do the new run on a NextSeq-2000.