Hi,

I've noticed some differences in the oligonucleotide sequences for SPLiT-seq between different protocol versions.

This is the structure of read2 based on the preprint, which matches what you currently have:
[10-bp UMI][8-bp Round3 barcode]GTGGCCGATGTTTCGCATCGGCGTACGACT[8-bp Round2 barcode]ATCCACGTGCTTGAGCGCGCTGCATACTTG[8-bp Round1 barcode]CCCATGATCGTCCGAAGGCCAGAGCATTCG(dT)

This is the structure of read2 based on the published manuscript:
[10-bp UMI][8-bp Round3 barcode]GTGGCCGATGTTTCGCATCGGCGTACGACT[8-bp Round2 barcode]ATCCACGTGCTTGAGAGGCCAGAGCATTCG[8-bp Round1 barcode](dT or N6). This sequence ends with either poly(dT)VN or NNNNNN depending on if the oligo primes off poly(A) or random hexamers. I don't know why the oligo design changed between preprint and acceptance.

Finally, the oligo design has changed again in the commercial version of SPLiT-seq by Parse BioSciences. This is the new read2 structure:
[10-bp UMI][8-bp Round3 barcode]GTGGCCGATGTTTCGCATCGGCGTACGACT[8-bp Round2 barcode]ATCCACGTGCTTGAGACTGTGG[8-bp Round1 barcode](dT)

As far as I can tell, all three versions have a 6 bp i7 index between the Illumina TruSeq and P7 sequences.