Different genome references

[ilyakorsunsky]

Thanks for this very timely resource! It looks like in the benchmarks of 10X pbmcs, one of the datasets was quantified with hg19 but the others with GRCh38. Is this intentional to integrate across different references? In a similar benchmark in the Harmony paper, I remapped everything to GRCh38. Not surprisingly, it made integration much easier

[wikiselev]

Interesting, thanks for the insight! We have not looked into that yet, for ease, just downloaded the datasets from the 10X website.