Xenograft Sorting:

The core of Xenomake is the implementation of xenograft sorting on reads that are mapped to both mouse and human organisms (overlapped reads). Xenograft Sorting is performed using Xengsort [¹¹], a faster and lightweight implementation than previous methods such as Xenome [¹⁰]. Xengsort is an alignment-free method that stores k-mers into a large table that enables the lookup of species of origin (host,graft, or both) of each RNA k-mer that occurs in either genome assembly. Reads are classified by iteratively searching k-mers for the appropriate species of origin. This process assigns reads into “host”, “graft”, “ambiguous”, “both”, or “neither” based upon Xengsort’s decision rule. When building the index files used in classification, Xengsort recommends storing 4.5 billion 25-mers when using humans and mouse genomes. After index generation, we implement Xengsort classify on our reads using default settings.