question about "Barcode and cDNA on the same mate"
houruiyan opened this issue · comments
Hi Alex,
I observed a phenonome when I run the new release function Barcode and cDNA on the same mate. You give us an example to set the readFilesIn. reads1 fastq is at the first palce and the reads2 fastq is at the second place. I found that the strand of reads is wrong when I run STARsolo according to this order.
Actually I check the reads strand by using is.reverse in pysam. There is an example (GAPDH).
This is the igv plot of GAPDH. It located at the forward strand.
When I run STARsolo according to reads1.fastq reads2.fastq and then fetch the reads during 6534511 to 6538374 in the chromosome 12.
Most of reads is reverse.
However when I changed the order of reads1.fastq and reads2.fastq, most of reads is not reverse.
Hi @houruiyan
The strand of the reads depends on the library protocol, STAR just maps the sequence to the proper strand.
If you change the order of the reads, you can indeed effectively reverse the mapped strands.
Alternatively, you can use --soloStrand to specify the strand for counting, but this does not affect the strand in BAM output.
Hi alex,
Actually I am confused about that, we used the 10x 3'scRNA-seq. When I used the order : reads2 reads1 rather than the reads1 reads2 and then I filtered bam file by using "GX" tag, I can get more filtered bam file. The reads2 reads1 strategy can get around 12G files, but the reads1 reads2 strategy can only get around 100M files.
So you mean for 10x genomics 5' data, the order should be reads1 and reads2 for readFilesIn when I use the --soloBarcodeMate to align the cDNA in cellbarcode. I should used the reads2 and reads1 for readFilesIn when I use the --soloBarcodeMate to align the cDNA in cellbarcode in 10x 3'scRNA-seq?
Hi @houruiyan
for 10X 3', you have to use read2 read1 order with the default --soloStrand Forward. The GX tag in the BAM files depends on the strand.
Hi alex,
Thank you very much for your reply! It is very useful!
Ruiyan