Abstract: Author summary TIPP3 is a new method for abundance profiling based on marker genes (i.e., genes that are universal and single copy) that improves on its predecessors, TIPP and TIPP2. By using advances in multiple sequence alignment and phylogenetic placement, TIPP3 is able to work with substantially larger reference packages (i.e., multiple sequence alignments and taxonomies for each of its marker genes) than in TIPP2, so that these contain ∼ 55,000 sequences per marker gene. This extension to large reference packages produces improved accuracy in abundance profiling compared to TIPP2 and also provides advantages over the current leading abundance profiling methods. Specifically, TIPP3 achieves higher accuracy than competing methods when input reads have high sequencing errors or are from novel genomes. The fast version, TIPP3-fast, is slightly less accurate than TIPP3 but nevertheless useful: compared to the most accurate competing methods, it is competitive in runtime and more accurate under challenging conditions. Finally, we show improvements to Kraken and Bracken when the input reads are filtered to those drawn from the marker genes. TIPP3 is open-source and available at https://github.com/c5shen/TIPP3.
Loading