We are delighted to introduce hera-T, a fast and accurate algorithm to quantify gene abundances from 10X-Chromium data with high rates of non-exonic reads, developed by BioTuring team.
hera-T devises a new strategy for aligning reads to both transcriptome and genome references that considerably reduces both the running time and memory consumption. For a single-cell data set with 49M reads, hera-T took only 1.75 minutes to process and consumed 8GB RAM. For a sample dataset of 11 billion reads (about 1.3 million cells), hera-T needs 24GB RAM and takes 4 hours to finish.
The t-SNE plots from hera-T and CellRanger are almost indistinguishable. Minor differences come from some difficult splicing alignment scenarios (with short exons) that hera-T addresses but Cell Ranger fails to address. (https://blog.bioturing.com/2018/10/12/cell-ranger-problems-and-hera-t/)
Please find some initial benchmarks of hera-T here.
To get hera-T package, please contact us at email@example.com.