Benchmarking dataset for measurement and removal of index hopping artifacts in multiplexed droplet-based single-cell RNA-seq data

Sample index hopping refers to the incorrect sample assignment of a demultiplexed sequencing read in a library pool. To enable benchmarking of methods for measurement of index hopping rate and removal of its artifacts in single-cell RNA-seq data, we developed a validation dataset consisting of a multiplexed library of two samples, in which the true sample of origin of most reads are known. The reads with known sample of origin provide the ground truth for measuring the performance of index hopping correcting methods. Two 10X Genomics scRNA-seq sample libraries were sequenced in two conditions. In the first condition, the samples were multiplexed on the same lane. In the second condition, two sample libraries were sequenced on two separate lanes of HiSeq 4000 (this non-multiplexed condition provides the ground truth for the true sample of origin of each read, based on the cell-UMI-gene information encoded in each read).