So, essentially the problem is that replicates are needed to estimate biological/technical error. Once you have this, “bulk” methods are very good. However, with commonly used “droplet” methods, replicates are expensive and prone to batch effects. But SPLiT-seq is multiplexable.