My lab has launched our first 10x GEX 5' libraries (v1.0). Cell viability was good (96%) and cDNA/SI-PCR product traces matched QC. Our strategy was downsample the libraries after over-sequencing to determine optimal depth for the cell type ( and aid in detecting the transgene). BCL's were converted to fastq via CRv.5.0.1. A custom reference was generated using the same CR. Our web summary html from CR count showed we had greater than expected cells (23K; 17.5K loaded expecting 10K) and a low median gene count of 131 genes.
Prior to launching CR count, I downsampled the fastq's using seqkt sample. With matching random seeds for R1/R2. I am wondering if this has contributed to the skewed cell/gene count?
Previously, our lab has used CR 3 to analyze 5' v1.0 data.
Right now I am consider rerunning with CR3 and pushing the full fastq (800million reads) through CR and downsampling afterwards.
10x tech support suggested a cell prep/lib prep issue either poor quality cells, GEM emulsion failure or premature lysis.
We did not sequence libraries in channels 7 & 8.
Any thoughts, advice would be appreciated. Thanks.