Compiling the sequence databases for your study is a critical task,
but analysis scenarios are too diverse to give an all-embracing
recipe. In the following we provide some general guidelines.- Make sure that your data collection is not a black box. You
should always know exactly which species are represented and
how they are related. Moreover, the assumption that the entire
gene set of a species is represented in the database should—at
least approximately—hold. Don’t forget in this context that the
non-detection of a pathway component in a particular species
might have a relevant functional implication (seeNote 4). - Select species such that they cover the full phylogenetic range
you are interested in. For example, if you are focusing on
pathway evolution in fungi, use a representative set of fungal
species plus a couple of close outgroup species, e.g., animals.
Plant or bacterial genomes are probably only adding to the
complexity of the analysis and are likely to provide no further
insights. - Sub-select species/strains such that you arrive at a desired level
of resolution in your analysis. For example, it does make little
sense to consider data from 25 different yeast strains if you are
interested in evolutionary events that happened a long time
before yeast existed as a species. However, at the same time,
Fig. 3Retrieving annotated pathway components from KEGG pathways database. (a) An example output from
clicking “KEGG GENES” link for the AMPK pathway. (b) Amino acid sequence for gene id “hsa:10000” obtained
from the KEGG database in FASTA format. The first line starting with “>” is the sequence identifier. (c)
Tab-delimited cross-reference table resulting from the online conversion of KEGG gene ids to UniProt ids
Tracing AMPK Evolution 119