Fig. 1

Schematic workflow for phylogenomic analysis and reconstruction of nuclear rRNA genes from single-cell transcriptomic data. (A) Initial gene trees were constructed using the taxon/gene-rich pipeline PhyloToL. Each tree was inspected, and subclades containing the maximum number of target taxa were extracted. Subclades were manually checked for non-target taxa. Curated gene trees with single-copy genes from all taxa were used to prepare the data matrix. (B) For concatenation analysis, each dataset with well-curated sequences were aligned and 265 datasets were concatenated into an amino acid supermatrix for maximum likelihood (ML) phylogeny. The ASTRAL method used 265 protein-coding gene trees to infer a species tree. Species delimitation and species tree inference were performed using BPP with 265 well-aligned datasets of nucleotides (C) Nuclear rRNA gene sequences were reconstructed from raw single-cell RNA sequencing (scRNA-seq) data using read mapping software (HISAT2 and Bowtie 2). Reconstructed sequences were aligned, and rRNA gene phylogeny was inferred using the ML approach. Detailed methods are provided in the Materials and methods section