Codon 1, 2 and 1+2 sequences were extracted from CDS alignments and used as input for building trees, along with protein and CDS sequences.
Representative sequences were extracted from each mPing subtype and aligned with four canonical defined mPing subtypes (mPingA, mPingB, mPingC, and mPingD) from the prior study5 using MUSCLE v3.8.42545 with default parameters (-maxiters 16).
Four sequences were extracted from GivenImaging capsule videos [4].
Candidate sequences were extracted from the 50 most highly upregulated genes from the microarray analyses.
Only continuous sequences were extracted from genomes.
Individual protein sequences were extracted from SWISSPROT database (release 55.0) (http://www.SwissProt.org) using protein IDs.
Spacer sequences were extracted from both genomes using perl script software.
8,998 paired SK/RR sequences were extracted from 769 fully sequenced bacterial genomes (see Methods).
The 16 T. thynnus sequences were extracted from the study by Alvarado Bremer et al. [15].
Sets of sequences were extracted from each of the all_mrna, stsMap, and all_est annotation tables.
All human coding sequences were extracted from RefSeq mRNA database at NCBI (27,184 sequences).
