Beyond classification: gene-family phylogenies from shotgun metagenomic reads enable accurate community analysis

BMC Genomics

Table 1 Gene families used in simulations

Gene name	Median length	Rate of amino acid evolution	Number of sequences	Source of sequences	Source of profile
16S rRNA	1535 bp	NA (Highly conserved)	427	RDP	RDP (INFERNAL)
rpoB	1296 aa	73.51	460	AMPHORA + GenBank	AMPHORA (HMMER)
rpsB	226 aa	51.96	411	AMPHORA + GenBank	AMPHORA (HMMER)
dnaG	395 aa	112.53	456	AMPHORA + GenBank	AMPHORA (HMMER)
lolC	411 aa	184.04	442	UniProt + GenBank	PhyloFacts (HMMER)

Each family of gene sequences was limited to its unique representatives among AMPHORA taxa (see Methods). Rate of amino acid evolution was determined by summing all branch lengths in a phylogenetic tree inferred via RAxML from the protein sequences; smaller values indicate fewer substitutions and greater conservation. The 16S rRNA gene requires a nucleotide model of evolution and hence has an incomparable value; it is well known to be highly conserved, with variable regions. 16S rRNA sequences were obtained from the Ribosomal Database Project (RDP) [20]. A larger set of 1,071 16S rRNA sequences was used only for the Fast UniFrac analysis (see Additional file 1: Table S1). Amino acid sequences for rpoB, rpsB, and dnaG families were obtained via AMPHORA [14], while corresponding DNA sequences were downloaded from NCBI GenBank [21]. For lolC, family members were determined by PhyloFacts [22] (family accession bpg052966 as of February 16, 2011); amino acid sequences were downloaded from UniProt [23], and corresponding DNA sequences were downloaded from EMBL-EBI [24]. Additional file 1: Table S1 provides download dates and sequence accession numbers.

ISSN: 1471-2164