Figure 1.
Correspondence between functional annotations in IMG using an example of ribosomal protein L1 and a combination of Pfam, COG, TIGRfam, FIGfam, IMG term and KO term annotations.
The same gene may be assigned to one or more functions according to each classification system; COG, Pfam and KO make no distinction between the cytosolic and organelle forms of ribosomal protein L1, while FIGfams include more than 10 synonymous annotations for proteins with different subcellular localization, taxonomic affiliation, and subsystem membership, such as “LSU ribosomal protein L1”, “LSU ribosomal protein L1p (L10Ae)” and “Ribosomal protein L1”. The sets of proteins classified as ribosomal protein L1 by each annotation system are represented by circles of different color; the size and extent of overlap between the circles reflect the relationship between sets of proteins annotated by each protein classification system.
Figure 2.
IMG tools for analysis of the correspondence between KO Terms and other protein families in IMG.
(i) The summary table listing all KO terms and their correspondence to other protein families (COGs, Pfams, TIGRfams) based on the genes assigned to both; for instance, KO term K00005 (glycerol dehydrogenase) corresponds to one COG, 2 Pfams and 0 TIGRfams, with 4 unique combinations of all 3 functional assignments. (ii) Detailed lists of genes corresponding to unique combinations of functional assignments using an example of K00005; 4 unique combinations of COG and Pfam for K00005 are shown with the counts of genes assigned to this particular combination, as well as the counts of genes with this combination of COG and Pfam and different KO term assignment or no KO term assignment. In this case the majority of the genes with K00005 term are assigned to COG0371 and pfam01761; however, many genes with the same COG and Pfam are assigned to other KO terms indicating either the lack of consistency of KO term assignment or the broadness of the corresponding COG and Pfam, which may include multiple paralogous proteins with different enzymatic activities.
Figure 3.
KO term distribution across genomes and paralog clusters in IMG.
(i) The summary table listing all KO terms with the number of genes to which they were assigned and the number of genomes in which these genes were found, including the average number of genes with this term per genome. The higher average number of genes with the term per genome and higher number of paralogs assigned with the same term can be indicative of over-annotation. (ii) Detailed view of the paralogous genes assigned with the same KO term; as an example, 2 genes in Saccharomyces cerevisiae genome are assigned with the term K00004, (R,R)-butanediol dehydrogenase. One of them appears to be an experimentally verified (2R,3R)-2,3-butanediol dehydrogenase, while the other protein is a putative polyol dehydrogenase with unknown specificity.
Figure 4.
IMG Pathway Definitions and Assertions.
(i) IMG Pathway Details displays information on a specific IMG pathway including the author and editing history, as well as the component reactions and functional roles. Assertion details for an IMG Pathway can be examined via (ii) the list of genomes for which the pathway is asserted, and (iii) the number of genes in the genome assigned to the functional roles in the pathway. Genomes can be compared in terms of IMG pathways using Function Profile analysis tool, with the result (iv) displaying counts of genes assigned to functional roles in the pathway.
Figure 5.
Phenotype Prediction Rules in IMG are defined as AND-OR rules based on IMG pathway assertions.
Phenotype rules range from complex, such as (i) the rule for predicting “Growth on cellulose via cellobiose” which involves multiple IMG pathways, to very simple, such as (ii) the rule for predicting “L-histidine prototroph” which involves one IMG pathway.
Table 1.
NOT (3-valued logic).
Table 2.
OR (3-valued logic).
Table 3.
AND (3-valued logic).
Figure 6.
Examining phenotype prediction results in IMG.
(i) All organisms that are predicted to be aerobic can be examined using a phylogenetic tree display; by clicking on “(?)”, (ii) users can examine the evidence of phenotype prediction of each organism.
Figure 7.
Metadata search and display in IMG.
(i) Metadata search with condition “aerobe with mesophile temperature range” returns (ii) a list of genomes that satisfy this condition; following the genome name hyperlink (iii) users can further examine organism detail and associated metadata information.