View original ARBA rule on UniProt
This rule predicts GO:0051998 (protein carboxyl O-methyltransferase activity) using 5 condition sets that identify two mechanistically distinct enzyme families: isoprenylcysteine carboxyl methyltransferases (ICMT) and protein-L-isoaspartate methyltransferases (PIMT). These enzymes share a common chemistry (methylating carboxyl groups on proteins) but differ fundamentally in substrate specificity, cellular location, biological function, and mechanism.
Condition-set counts describe the sets recorded in this review, which may omit the full rule.
Interactive prediction matrix showing how row entries PREDICT column entries. Cell (i,j) shows what fraction of proteins with row domain i also have column domain j. Click cells to view intersection in UniProt. Click domain IDs to view proteins with that domain.
Legend: Each cell shows PREDICTS % (fraction of row entry proteins that also have column entry - row PREDICTS column), Jaccard similarity (J:%), and intersection count. CS = Condition Set(s), TGT = GO annotation target.
ARBA00026372 exhibits severe structural flaws that undermine its biological validity. The rule conflates two mechanistically distinct enzyme families - isoprenylcysteine carboxyl methyltransferases (ICMT) and protein-L-isoaspartate methyltransferases (PIMT) - under a single overly broad GO term. While both enzymes methylate carboxyl groups on proteins, they differ fundamentally in substrate specificity, cellular location, biological function, and mechanism. ICMT enzymes are integral ER membrane proteins that modify prenylated C-terminal cysteines to enhance membrane affinity of signaling proteins like Ras. PIMT enzymes are cytosolic repair enzymes that methylate isoaspartyl residues in damaged proteins as part of a repair cycle. The complete DISJOINT nature of condition sets identifying these two families (Jaccard=0.000 between CS1 and CS2, CS1 and CS5) confirms they represent separate protein families that should not be annotated with the same specific GO term. Additionally, condition set 1 contains internal redundancy (IPR025770 is a complete subset of IPR007269), and the rule uses the generic parent term GO:0051998 when specific child terms GO:0004671 (for ICMT) and GO:0004719 (for PIMT) exist. The average Jaccard similarity of 0.097 across domain pairs reflects the fundamental heterogeneity of this rule.
This rule should be REMOVED and replaced with two separate rules, each using appropriate specific GO terms. The conflation of ICMT and PIMT enzymes under GO:0051998 represents a fundamental violation of the principle that automated rules should capture proteins with shared mechanistic and functional characteristics. While both enzyme families perform carboxyl O-methylation, this shared chemistry is insufficient justification for a single annotation rule given their profound differences in substrate, mechanism, cellular context, and biological role. The complete protein-level disjointness between ICMT and PIMT condition sets (0% overlap) definitively demonstrates these are distinct enzyme families that evolution has specialized for different purposes. Using a broad parent GO term to paper over mechanistic differences creates misleading annotations that obscure rather than clarify protein function. Furthermore, the internal redundancy within condition set 1 (IPR025770 completely contained in IPR007269) indicates poor rule design. This rule should be deprecated and replaced with two parsimonious rules: one annotating ICMT proteins to GO:0004671 (protein C-terminal S-isoprenylcysteine carboxyl O-methyltransferase activity) and one annotating PIMT proteins to GO:0004719 (protein-L-isoaspartate (D-aspartate) O-methyltransferase activity).
Condition set 1 identifies ICMT enzymes through InterPro and PANTHER signatures. These are integral ER membrane proteins (8 transmembrane domains) that methylate prenylated cysteine residues at the C-terminus of CaaX proteins. IPR025770 is completely contained within IPR007269 (12 of 12 proteins overlap, Jaccard=0.750), indicating redundancy within this condition set.
Condition set 2 identifies PIMT enzymes in eukaryotes. These are cytosolic protein repair enzymes that methylate isoaspartyl residues in damaged proteins. This condition set is completely DISJOINT from condition set 1 (0 protein overlap), representing a mechanistically distinct enzyme family.
Condition set 3 identifies additional ICMT-family proteins (6 proteins in SwissProt). All 6 proteins are subsets of the IPR007269 and IPR025770 sets from CS1, but this FunFam captures only a portion of the ICMT family. No taxonomic restriction is applied.
Condition set 4 identifies fungal ICMT proteins (2 proteins in SwissProt). These 2 proteins are subsets of IPR025770 and IPR007269 from CS1, but are DISJOINT from the 6 proteins in CS3 (1.20.120.1630:FF:000007), suggesting these FunFams partition the ICMT family.
Condition set 5 identifies metazoan PIMT proteins (2 proteins in SwissProt). This condition set is completely DISJOINT from all ICMT-related condition sets (CS1, CS3, CS4) and also DISJOINT from the eukaryotic PIMT set in CS2, indicating these FunFams partition the PIMT family by taxonomy.
The rule contains 5 condition sets attempting to capture two distinct enzyme families, with significant redundancy and unnecessary complexity. Condition set 1 alone contains redundant InterPro domains (IPR025770 is a complete subset of IPR007269 with 100% containment). The inclusion of multiple taxonomically restricted FunFam condition sets (CS2 for Eukaryota, CS4 for Fungi, CS5 for Metazoa) adds complexity without biological justification when these could be consolidated within family-specific rules. The fundamental issue is not complexity per se, but rather that the rule attempts to unify mechanistically distinct enzyme families that should be handled by separate rules. The average Jaccard similarity of 0.097 across all domain pairs reflects this inappropriate grouping rather than acceptable biological diversity within a single enzyme family.
Literature clearly establishes that ICMT and PIMT are mechanistically distinct enzyme families despite sharing a common chemistry. ICMT is an integral ER membrane protein with 8 transmembrane domains that methylates the alpha-carboxyl group of prenylated cysteine residues at the C-terminus of CaaX proteins. This modification is permanent and increases membrane affinity of signaling proteins like Ras. In contrast, PIMT is a cytosolic enzyme that methylates the alpha-carboxyl group of isoaspartyl residues in damaged or aging proteins. This methylation is transient and initiates a repair cycle through a succinimide intermediate. The substrates differ (prenylated C-terminal cysteine vs. internal isoaspartyl residues), the cellular locations differ (ER membrane vs. cytosol), the biological functions differ (signal protein modification vs. protein repair), and the outcomes differ (permanent membrane targeting vs. transient repair intermediate). These fundamental differences contradict the premise that these enzymes should receive identical GO annotations simply because they both methylate carboxyl groups. The literature supports using specific child GO terms that capture these mechanistic distinctions.
The rule exhibits severe overlap issues at multiple levels. Within condition set 1, IPR025770 is completely contained within IPR007269 (100% containment, 75% Jaccard), indicating unnecessary redundancy. More critically, the rule shows systematic DISJOINT relationships between major condition set groups. Condition sets identifying ICMT enzymes (CS1, CS3, CS4) are completely DISJOINT from condition sets identifying PIMT enzymes (CS2, CS5), with 0% protein overlap. For example, IPR007269 (CS1) vs. 3.40.50.150:FF:000027 (CS2) shows Jaccard=0.000 with 0 shared proteins out of 28 total. Similarly, IPR025770 (CS1) vs. 3.40.50.150:FF:000235 (CS5) shows Jaccard=0.000 with 0 shared proteins out of 14 total. Even within enzyme families, FunFam condition sets show suspicious disjointness: 1.20.120.1630:FF:000007 (CS3) and 1.20.120.1630:FF:000018 (CS4) are completely DISJOINT despite both being labeled as protein-S-isoprenylcysteine O-methyltransferase. The systematic pattern of DISJOINT relationships (9 out of 15 domain pairs show Jaccard=0.000) indicates this rule groups unrelated protein families rather than capturing a single coherent enzyme family with some structural diversity.
GO:0051998 (protein carboxyl O-methyltransferase activity) is a generic parent term that obscures critical mechanistic and functional differences between the enzyme families captured by this rule. The GO hierarchy includes specific child terms that accurately reflect the distinct substrate specificities and biological roles of these enzymes: GO:0004671 (protein C-terminal S-isoprenylcysteine carboxyl O-methyltransferase activity) for ICMT enzymes and GO:0004719 (protein-L-isoaspartate (D-aspartate) O-methyltransferase activity) for PIMT enzymes. Using the broad parent term GO:0051998 for both enzyme families is analogous to annotating all kinases with "protein kinase activity" despite their diverse substrate specificities and regulatory roles. The choice of this broad term appears to be a workaround to justify combining mechanistically unrelated enzymes in a single rule, rather than a biologically appropriate annotation level. Proper curation practice would split this rule and use the specific child terms that capture the distinct mechanistic features of ICMT vs. PIMT enzymes. The definition of GO:0051998 ("Catalysis of the transfer of a methyl group to a carboxyl group on a protein") is accurate but insufficiently specific to be informative given that specific child terms exist for the different target residues and contexts.
The taxonomic restrictions in this rule are inconsistent and incoherent. Condition set 1 (ICMT enzymes via InterPro domains) lacks any taxonomic restriction and would annotate all domains of life, while condition set 4 (also ICMT enzymes but via a different FunFam) is restricted to Fungi (NCBITaxon:4751). This creates an illogical situation where the same enzyme family receives universal annotation in one condition set but fungi-specific annotation in another. Similarly, condition set 2 restricts PIMT enzymes to Eukaryota (NCBITaxon:2759) while condition set 5 restricts PIMT enzymes to Metazoa (NCBITaxon:33208), creating overlapping redundant taxonomic coverage. The fundamental issue is that this multi-family rule conflates two distinct enzyme families that likely have different phylogenetic distributions. ICMT enzymes appear to be primarily eukaryotic (involved in CaaX protein processing), while PIMT enzymes are more broadly distributed across bacteria and eukaryotes. Condition set 1 is MISSING an appropriate taxonomic restriction for ICMT, which should likely be restricted to Eukaryota. If the rule were properly split into separate ICMT and PIMT rules, each could have internally consistent taxonomic restrictions based on the actual phylogenetic distribution of each enzyme family.
IPR025770 is a complete subset of IPR007269 with 100% containment
Average Jaccard similarity across all domain pairs is 0.097
9 out of 15 domain pairs show complete disjointness (Jaccard=0.000)
ICMT-related domains (IPR007269, IPR025770) are completely DISJOINT from PIMT-related domains (3.40.50.150:FF:000027, 3.40.50.150:FF:000235)
Rule contains 5 condition sets with significant structural issues
Only 1 pair shows >50% overlap (the redundant IPR007269/IPR025770 pair within CS1)
Rule annotates to GO:0051998 (protein carboxyl O-methyltransferase activity)
Rule created 2021-10-20, last modified 2025-03-21
Condition sets mix ICMT and PIMT enzyme families with inconsistent taxonomic restrictions
ICMT is an integral ER membrane protein with 8 transmembrane domains
ICMT methylates prenylated cysteine residues at C-terminus of CaaX proteins
Methylation increases membrane affinity by neutralizing negative charge
PIMT is a cytosolic repair enzyme for damaged proteins
PIMT methylates isoaspartyl residues formed from asparagine deamidation or aspartate dehydration
Methylation is transient and initiates repair cycle through succinimide intermediate
ICMT performs the third step of CaaX protein processing after prenylation and proteolysis
Modification is permanent and enhances membrane localization of signaling proteins
PIMT is highly conserved across eukaryotes, archaebacteria, and Gram-negative bacteria
Enzyme recognizes and repairs damaged L-isoaspartyl and D-aspartyl residues in aging proteins
id: ARBA00026372
description: 'This rule predicts GO:0051998 (protein carboxyl O-methyltransferase
activity) using 5 condition sets that identify two mechanistically distinct enzyme
families: isoprenylcysteine carboxyl methyltransferases (ICMT) and protein-L-isoaspartate
methyltransferases (PIMT). These enzymes share a common chemistry (methylating carboxyl
groups on proteins) but differ fundamentally in substrate specificity, cellular
location, biological function, and mechanism.'
status: COMPLETE
rule_type: ARBA
rule:
rule_id: ARBA00026372
condition_sets:
- number: 1
conditions:
- condition_type: INTERPRO
value: IPR007269
curie: InterPro:IPR007269
label: Isoprenylcysteine carboxyl methyltransferase
negated: false
- condition_type: INTERPRO
value: IPR025770
curie: InterPro:IPR025770
label: Protein-S-isoprenylcysteine O-methyltransferase
negated: false
- condition_type: PANTHER
value: PTHR12714
curie: PTHR12714
label: ''
negated: false
notes: Condition set 1 identifies ICMT enzymes through InterPro and PANTHER signatures.
These are integral ER membrane proteins (8 transmembrane domains) that methylate
prenylated cysteine residues at the C-terminus of CaaX proteins. IPR025770 is
completely contained within IPR007269 (12 of 12 proteins overlap, Jaccard=0.750),
indicating redundancy within this condition set.
pairwise_overlap:
- condition_a: IPR007269
condition_b: IPR025770
protein_database: SWISSPROT
count_a: 16
count_b: 12
intersection_count: 12
a_minus_b_count: 4
b_minus_a_count: 0
jaccard_similarity: 0.75
containment_a_in_b: 0.75
containment_b_in_a: 1.0
interpretation: SUBSET
- condition_a: IPR007269
condition_b: PTHR12714
protein_database: SWISSPROT
count_a: 16
count_b: 19
intersection_count: 14
a_minus_b_count: 2
b_minus_a_count: 5
jaccard_similarity: 0.6666666666666666
containment_a_in_b: 0.875
containment_b_in_a: 0.7368421052631579
interpretation: HIGH_OVERLAP
- condition_a: IPR025770
condition_b: PTHR12714
protein_database: SWISSPROT
count_a: 12
count_b: 19
intersection_count: 12
a_minus_b_count: 0
b_minus_a_count: 7
jaccard_similarity: 0.631578947368421
containment_a_in_b: 1.0
containment_b_in_a: 0.631578947368421
interpretation: SUBSET
- number: 2
conditions:
- condition_type: FUNFAM
value: 3.40.50.150:FF:000027
curie: CATH.FunFam:3.40.50.150:FF:000027
label: Protein-L-isoaspartate O-methyltransferase
negated: false
- condition_type: TAXON
value: '2759'
curie: NCBITaxon:2759
label: Eukaryota
negated: false
notes: Condition set 2 identifies PIMT enzymes in eukaryotes. These are cytosolic
protein repair enzymes that methylate isoaspartyl residues in damaged proteins.
This condition set is completely DISJOINT from condition set 1 (0 protein overlap),
representing a mechanistically distinct enzyme family.
- number: 3
conditions:
- condition_type: FUNFAM
value: 1.20.120.1630:FF:000007
curie: CATH.FunFam:1.20.120.1630:FF:000007
label: Protein-S-isoprenylcysteine O-methyltransferase
negated: false
notes: Condition set 3 identifies additional ICMT-family proteins (6 proteins
in SwissProt). All 6 proteins are subsets of the IPR007269 and IPR025770 sets
from CS1, but this FunFam captures only a portion of the ICMT family. No taxonomic
restriction is applied.
- number: 4
conditions:
- condition_type: FUNFAM
value: 1.20.120.1630:FF:000018
curie: CATH.FunFam:1.20.120.1630:FF:000018
label: Protein-S-isoprenylcysteine O-methyltransferase
negated: false
- condition_type: TAXON
value: '4751'
curie: NCBITaxon:4751
label: Fungi
negated: false
notes: Condition set 4 identifies fungal ICMT proteins (2 proteins in SwissProt).
These 2 proteins are subsets of IPR025770 and IPR007269 from CS1, but are DISJOINT
from the 6 proteins in CS3 (1.20.120.1630:FF:000007), suggesting these FunFams
partition the ICMT family.
- number: 5
conditions:
- condition_type: FUNFAM
value: 3.40.50.150:FF:000235
curie: CATH.FunFam:3.40.50.150:FF:000235
label: Protein-L-isoaspartate O-methyltransferase
negated: false
- condition_type: TAXON
value: '33208'
curie: NCBITaxon:33208
label: Metazoa
negated: false
notes: Condition set 5 identifies metazoan PIMT proteins (2 proteins in SwissProt).
This condition set is completely DISJOINT from all ICMT-related condition sets
(CS1, CS3, CS4) and also DISJOINT from the eukaryotic PIMT set in CS2, indicating
these FunFams partition the PIMT family by taxonomy.
go_annotations:
- go_id: GO:0051998
go_label: protein carboxyl O-methyltransferase activity
aspect: MF
entries:
- id: 1.20.120.1630:FF:000007
type: FUNFAM
label: Protein-S-isoprenylcysteine O-methyltransferase
appears_in_condition_sets:
- 3
protein_count: 6
related_entries:
- relationship: PREDICTS
target_id: IPR007269
containment: 0.375
jaccard_similarity: 0.375
intersection_count: 6
exclusive_count: 10
- relationship: PREDICTS
target_id: IPR025770
containment: 0.5
jaccard_similarity: 0.5
intersection_count: 6
exclusive_count: 6
- relationship: PREDICTS
target_id: PTHR12714
containment: 0.316
jaccard_similarity: 0.316
intersection_count: 6
exclusive_count: 13
- relationship: EQUIV
target_id: 3.40.50.150:FF:000027
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 6
- relationship: EQUIV
target_id: 1.20.120.1630:FF:000018
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 6
- relationship: EQUIV
target_id: 3.40.50.150:FF:000235
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 6
- relationship: PREDICTS
target_id: GO:0051998
containment: 1.0
jaccard_similarity: 0.015
intersection_count: 6
exclusive_count: 0
- id: 1.20.120.1630:FF:000018
type: FUNFAM
label: Protein-S-isoprenylcysteine O-methyltransferase
appears_in_condition_sets:
- 4
protein_count: 2
related_entries:
- relationship: PREDICTS
target_id: IPR007269
containment: 0.125
jaccard_similarity: 0.125
intersection_count: 2
exclusive_count: 14
- relationship: PREDICTS
target_id: IPR025770
containment: 0.167
jaccard_similarity: 0.167
intersection_count: 2
exclusive_count: 10
- relationship: PREDICTS
target_id: PTHR12714
containment: 0.105
jaccard_similarity: 0.105
intersection_count: 2
exclusive_count: 17
- relationship: EQUIV
target_id: 3.40.50.150:FF:000027
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 2
- relationship: EQUIV
target_id: 1.20.120.1630:FF:000007
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 2
- relationship: EQUIV
target_id: 3.40.50.150:FF:000235
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 2
- relationship: PREDICTS
target_id: GO:0051998
containment: 1.0
jaccard_similarity: 0.005
intersection_count: 2
exclusive_count: 0
- id: 3.40.50.150:FF:000027
type: FUNFAM
label: Protein-L-isoaspartate O-methyltransferase
appears_in_condition_sets:
- 2
protein_count: 12
related_entries:
- relationship: EQUIV
target_id: IPR007269
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 12
- relationship: EQUIV
target_id: IPR025770
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 12
- relationship: EQUIV
target_id: PTHR12714
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 12
- relationship: EQUIV
target_id: 1.20.120.1630:FF:000007
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 12
- relationship: EQUIV
target_id: 1.20.120.1630:FF:000018
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 12
- relationship: EQUIV
target_id: 3.40.50.150:FF:000235
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 12
- relationship: PREDICTS
target_id: GO:0051998
containment: 1.0
jaccard_similarity: 0.029
intersection_count: 12
exclusive_count: 0
- id: 3.40.50.150:FF:000235
type: FUNFAM
label: Protein-L-isoaspartate O-methyltransferase
appears_in_condition_sets:
- 5
protein_count: 2
related_entries:
- relationship: EQUIV
target_id: IPR007269
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 2
- relationship: EQUIV
target_id: IPR025770
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 2
- relationship: EQUIV
target_id: PTHR12714
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 2
- relationship: EQUIV
target_id: 3.40.50.150:FF:000027
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 2
- relationship: EQUIV
target_id: 1.20.120.1630:FF:000007
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 2
- relationship: EQUIV
target_id: 1.20.120.1630:FF:000018
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 2
- relationship: PREDICTS
target_id: GO:0051998
containment: 1.0
jaccard_similarity: 0.005
intersection_count: 2
exclusive_count: 0
- id: IPR007269
type: INTERPRO
label: Isoprenylcysteine carboxyl methyltransferase
appears_in_condition_sets:
- 1
protein_count: 16
related_entries:
- relationship: PREDICTED_BY
target_id: IPR025770
containment: 1.0
jaccard_similarity: 0.75
intersection_count: 12
exclusive_count: 0
- relationship: PREDICTS
target_id: PTHR12714
containment: 0.875
jaccard_similarity: 0.667
intersection_count: 14
exclusive_count: 2
- relationship: EQUIV
target_id: 3.40.50.150:FF:000027
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 16
- relationship: PREDICTED_BY
target_id: 1.20.120.1630:FF:000007
containment: 1.0
jaccard_similarity: 0.375
intersection_count: 6
exclusive_count: 0
- relationship: PREDICTED_BY
target_id: 1.20.120.1630:FF:000018
containment: 1.0
jaccard_similarity: 0.125
intersection_count: 2
exclusive_count: 0
- relationship: EQUIV
target_id: 3.40.50.150:FF:000235
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 16
- relationship: PREDICTS
target_id: GO:0051998
containment: 1.0
jaccard_similarity: 0.039
intersection_count: 16
exclusive_count: 0
- id: IPR025770
type: INTERPRO
label: Protein-S-isoprenylcysteine O-methyltransferase
appears_in_condition_sets:
- 1
protein_count: 12
related_entries:
- relationship: PREDICTS
target_id: IPR007269
containment: 0.75
jaccard_similarity: 0.75
intersection_count: 12
exclusive_count: 4
- relationship: PREDICTS
target_id: PTHR12714
containment: 1.0
jaccard_similarity: 0.632
intersection_count: 12
exclusive_count: 0
- relationship: EQUIV
target_id: 3.40.50.150:FF:000027
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 12
- relationship: PREDICTED_BY
target_id: 1.20.120.1630:FF:000007
containment: 1.0
jaccard_similarity: 0.5
intersection_count: 6
exclusive_count: 0
- relationship: PREDICTED_BY
target_id: 1.20.120.1630:FF:000018
containment: 1.0
jaccard_similarity: 0.167
intersection_count: 2
exclusive_count: 0
- relationship: EQUIV
target_id: 3.40.50.150:FF:000235
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 12
- relationship: PREDICTS
target_id: GO:0051998
containment: 1.0
jaccard_similarity: 0.029
intersection_count: 12
exclusive_count: 0
- id: PTHR12714
type: PANTHER
appears_in_condition_sets:
- 1
protein_count: 19
related_entries:
- relationship: PREDICTED_BY
target_id: IPR007269
containment: 0.737
jaccard_similarity: 0.667
intersection_count: 14
exclusive_count: 5
- relationship: PREDICTED_BY
target_id: IPR025770
containment: 0.632
jaccard_similarity: 0.632
intersection_count: 12
exclusive_count: 7
- relationship: EQUIV
target_id: 3.40.50.150:FF:000027
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 19
- relationship: PREDICTED_BY
target_id: 1.20.120.1630:FF:000007
containment: 1.0
jaccard_similarity: 0.316
intersection_count: 6
exclusive_count: 0
- relationship: PREDICTED_BY
target_id: 1.20.120.1630:FF:000018
containment: 1.0
jaccard_similarity: 0.105
intersection_count: 2
exclusive_count: 0
- relationship: EQUIV
target_id: 3.40.50.150:FF:000235
containment: 0.0
jaccard_similarity: 0.0
intersection_count: 0
exclusive_count: 19
- relationship: PREDICTS
target_id: GO:0051998
containment: 0.737
jaccard_similarity: 0.034
intersection_count: 14
exclusive_count: 5
review_summary: ARBA00026372 exhibits severe structural flaws that undermine its biological
validity. The rule conflates two mechanistically distinct enzyme families - isoprenylcysteine
carboxyl methyltransferases (ICMT) and protein-L-isoaspartate methyltransferases
(PIMT) - under a single overly broad GO term. While both enzymes methylate carboxyl
groups on proteins, they differ fundamentally in substrate specificity, cellular
location, biological function, and mechanism. ICMT enzymes are integral ER membrane
proteins that modify prenylated C-terminal cysteines to enhance membrane affinity
of signaling proteins like Ras. PIMT enzymes are cytosolic repair enzymes that methylate
isoaspartyl residues in damaged proteins as part of a repair cycle. The complete
DISJOINT nature of condition sets identifying these two families (Jaccard=0.000
between CS1 and CS2, CS1 and CS5) confirms they represent separate protein families
that should not be annotated with the same specific GO term. Additionally, condition
set 1 contains internal redundancy (IPR025770 is a complete subset of IPR007269),
and the rule uses the generic parent term GO:0051998 when specific child terms GO:0004671
(for ICMT) and GO:0004719 (for PIMT) exist. The average Jaccard similarity of 0.097
across domain pairs reflects the fundamental heterogeneity of this rule.
action: DEPRECATE
action_rationale: 'This rule should be REMOVED and replaced with two separate rules,
each using appropriate specific GO terms. The conflation of ICMT and PIMT enzymes
under GO:0051998 represents a fundamental violation of the principle that automated
rules should capture proteins with shared mechanistic and functional characteristics.
While both enzyme families perform carboxyl O-methylation, this shared chemistry
is insufficient justification for a single annotation rule given their profound
differences in substrate, mechanism, cellular context, and biological role. The
complete protein-level disjointness between ICMT and PIMT condition sets (0% overlap)
definitively demonstrates these are distinct enzyme families that evolution has
specialized for different purposes. Using a broad parent GO term to paper over mechanistic
differences creates misleading annotations that obscure rather than clarify protein
function. Furthermore, the internal redundancy within condition set 1 (IPR025770
completely contained in IPR007269) indicates poor rule design. This rule should
be deprecated and replaced with two parsimonious rules: one annotating ICMT proteins
to GO:0004671 (protein C-terminal S-isoprenylcysteine carboxyl O-methyltransferase
activity) and one annotating PIMT proteins to GO:0004719 (protein-L-isoaspartate
(D-aspartate) O-methyltransferase activity).'
suggested_modifications:
- Create ARBA rule for ICMT family using InterPro:IPR007269, PANTHER:PTHR12714, and
relevant FunFams (1.20.120.1630:FF:000007, 1.20.120.1630:FF:000018), annotating
to GO:0004671 (protein C-terminal S-isoprenylcysteine carboxyl O-methyltransferase
activity)
- Create separate ARBA rule for PIMT family using FunFams 3.40.50.150:FF:000027 and
3.40.50.150:FF:000235 with appropriate taxonomic restrictions, annotating to GO:0004719
(protein-L-isoaspartate (D-aspartate) O-methyltransferase activity)
- Remove IPR025770 from ICMT rule as it is completely redundant with IPR007269
- Deprecate current ARBA00026372 rule to prevent future false annotations
parsimony:
assessment: OVERLY_COMPLEX
notes: The rule contains 5 condition sets attempting to capture two distinct enzyme
families, with significant redundancy and unnecessary complexity. Condition set
1 alone contains redundant InterPro domains (IPR025770 is a complete subset of
IPR007269 with 100% containment). The inclusion of multiple taxonomically restricted
FunFam condition sets (CS2 for Eukaryota, CS4 for Fungi, CS5 for Metazoa) adds
complexity without biological justification when these could be consolidated within
family-specific rules. The fundamental issue is not complexity per se, but rather
that the rule attempts to unify mechanistically distinct enzyme families that
should be handled by separate rules. The average Jaccard similarity of 0.097 across
all domain pairs reflects this inappropriate grouping rather than acceptable biological
diversity within a single enzyme family.
supported_by:
- reference_id: file:rules/arba/ARBA00026372/ARBA00026372-analysis.yaml
supporting_text: '"IPR007269 ↔ IPR025770: Jaccard similarity: 0.75, Containment:
A in B = 0.750, B in A = 1.000, Interpretation: SUBSET. Summary: Analyzed 15
domain-domain pairs and 6 domain-GO pairs across entire rule. Average Jaccard
similarity: 0.097. 1 pairs with >50% overlap, 11 subset relationships."'
literature_support:
assessment: CONTRADICTED
notes: Literature clearly establishes that ICMT and PIMT are mechanistically distinct
enzyme families despite sharing a common chemistry. ICMT is an integral ER membrane
protein with 8 transmembrane domains that methylates the alpha-carboxyl group
of prenylated cysteine residues at the C-terminus of CaaX proteins. This modification
is permanent and increases membrane affinity of signaling proteins like Ras. In
contrast, PIMT is a cytosolic enzyme that methylates the alpha-carboxyl group
of isoaspartyl residues in damaged or aging proteins. This methylation is transient
and initiates a repair cycle through a succinimide intermediate. The substrates
differ (prenylated C-terminal cysteine vs. internal isoaspartyl residues), the
cellular locations differ (ER membrane vs. cytosol), the biological functions
differ (signal protein modification vs. protein repair), and the outcomes differ
(permanent membrane targeting vs. transient repair intermediate). These fundamental
differences contradict the premise that these enzymes should receive identical
GO annotations simply because they both methylate carboxyl groups. The literature
supports using specific child GO terms that capture these mechanistic distinctions.
supported_by:
- reference_id: web:https://pubmed.ncbi.nlm.nih.gov/22195972/
supporting_text: '"ICMT is a highly conserved enzyme that methyl esterifies the
alpha carboxyl group of prenylated proteins including Ras and related GTPases.
Methyl esterification neutralizes the negative charge of the prenylcysteine
and thereby increases membrane affinity. Icmt is an integral membrane protein
restricted to the endoplasmic reticulum (ER). Human Icmt traverses the ER membrane
eight times, with both N and C termini disposed toward the cytosol."'
- reference_id: web:https://pubmed.ncbi.nlm.nih.gov/31234761/
supporting_text: '"PIMT acts to transfer methyl groups from S-adenosyl-L-methionine
to the alpha side chain carboxyl groups of damaged L-isoaspartyl and D-aspartyl
amino acids. Under physiological conditions, deamidation of asparagine residues
or dehydration of aspartic acid residues results in formation of a metastable
intermediate succinimide that spontaneously hydrolyzes to form a mixture of
normal L-aspartyl and atypical L-isoaspartyl linkages. PIMT, using AdoMet as
a methyl donor, selectively methylates the isoaspartyl α-carboxyl group to form
a highly labile methyl ester. Spontaneous demethylation occurs within minutes
to reform the original succinimide, and further cycles of repair result in near
complete conversion of the isoaspartyl β-linkages to normal aspartyl α-linkages."'
- reference_id: web:https://en.wikipedia.org/wiki/Protein-S-isoprenylcysteine_O-methyltransferase
supporting_text: '"A C-terminal CaaX motif targets eukaryotic proteins to posttranslational
modifications. Processing is initiated by covalent attachment of a 15-carbon
farnesyl or a 20-carbon geranylgeranyl lipid to the cysteine, followed by removal
of the C-terminal three amino acids by CaaX protease Rce1, and then methylation
of the prenylcysteine by ICMT."'
condition_overlap:
assessment: SIGNIFICANT
notes: 'The rule exhibits severe overlap issues at multiple levels. Within condition
set 1, IPR025770 is completely contained within IPR007269 (100% containment, 75%
Jaccard), indicating unnecessary redundancy. More critically, the rule shows systematic
DISJOINT relationships between major condition set groups. Condition sets identifying
ICMT enzymes (CS1, CS3, CS4) are completely DISJOINT from condition sets identifying
PIMT enzymes (CS2, CS5), with 0% protein overlap. For example, IPR007269 (CS1)
vs. 3.40.50.150:FF:000027 (CS2) shows Jaccard=0.000 with 0 shared proteins out
of 28 total. Similarly, IPR025770 (CS1) vs. 3.40.50.150:FF:000235 (CS5) shows
Jaccard=0.000 with 0 shared proteins out of 14 total. Even within enzyme families,
FunFam condition sets show suspicious disjointness: 1.20.120.1630:FF:000007 (CS3)
and 1.20.120.1630:FF:000018 (CS4) are completely DISJOINT despite both being labeled
as protein-S-isoprenylcysteine O-methyltransferase. The systematic pattern of
DISJOINT relationships (9 out of 15 domain pairs show Jaccard=0.000) indicates
this rule groups unrelated protein families rather than capturing a single coherent
enzyme family with some structural diversity.'
supported_by:
- reference_id: file:rules/arba/ARBA00026372/ARBA00026372-analysis.yaml
supporting_text: '"IPR007269 ↔ 3.40.50.150:FF:000027: Counts: 16 / 12, Intersection:
0, Jaccard similarity: 0.0, Containment: A in B = 0.0, B in A = 0.0, Interpretation:
DISJOINT. IPR025770 ↔ 3.40.50.150:FF:000027: Counts: 12 / 12, Intersection:
0, Jaccard similarity: 0.0, Interpretation: DISJOINT. IPR025770 ↔ 3.40.50.150:FF:000235:
Counts: 12 / 2, Intersection: 0, Jaccard similarity: 0.0, Interpretation: DISJOINT.
3.40.50.150:FF:000027 ↔ 1.20.120.1630:FF:000007: Counts: 12 / 6, Intersection:
0, Jaccard similarity: 0.0, Interpretation: DISJOINT. 1.20.120.1630:FF:000007
↔ 1.20.120.1630:FF:000018: Counts: 6 / 2, Intersection: 0, Jaccard similarity:
0.0, Interpretation: DISJOINT."'
- reference_id: file:rules/arba/ARBA00026372/ARBA00026372-analysis.txt
supporting_text: '"Summary: Analyzed 15 domain-domain pairs and 6 domain-GO pairs
across entire rule. Average Jaccard similarity: 0.097. 1 pairs with >50% overlap,
11 subset relationships."'
go_specificity:
assessment: TOO_BROAD
notes: 'GO:0051998 (protein carboxyl O-methyltransferase activity) is a generic
parent term that obscures critical mechanistic and functional differences between
the enzyme families captured by this rule. The GO hierarchy includes specific
child terms that accurately reflect the distinct substrate specificities and biological
roles of these enzymes: GO:0004671 (protein C-terminal S-isoprenylcysteine carboxyl
O-methyltransferase activity) for ICMT enzymes and GO:0004719 (protein-L-isoaspartate
(D-aspartate) O-methyltransferase activity) for PIMT enzymes. Using the broad
parent term GO:0051998 for both enzyme families is analogous to annotating all
kinases with "protein kinase activity" despite their diverse substrate specificities
and regulatory roles. The choice of this broad term appears to be a workaround
to justify combining mechanistically unrelated enzymes in a single rule, rather
than a biologically appropriate annotation level. Proper curation practice would
split this rule and use the specific child terms that capture the distinct mechanistic
features of ICMT vs. PIMT enzymes. The definition of GO:0051998 ("Catalysis of
the transfer of a methyl group to a carboxyl group on a protein") is accurate
but insufficiently specific to be informative given that specific child terms
exist for the different target residues and contexts.'
supported_by:
- reference_id: file:rules/arba/ARBA00026372/ARBA00026372.enriched.json
supporting_text: '"annotations: [{annotationType: ANNOTATION, dbReference: {database:
GO, id: GO:0051998, properties: [{key: GoTerm, value: -}, {key: GoEvidenceType,
value: :-}], label: protein carboxyl O-methyltransferase activity}}]"'
- reference_id: web:GO_term_search
supporting_text: '"GO:0004671 (protein C-terminal S-isoprenylcysteine carboxyl
O-methyltransferase activity): Catalysis of the reaction: S-adenosyl-L-methionine
+ protein C-terminal S-farnesyl-L-cysteine = S-adenosyl-L-homocysteine + protein
C-terminal S-farnesyl-L-cysteine methyl ester. GO:0004719 (protein-L-isoaspartate
(D-aspartate) O-methyltransferase activity): Catalysis of the reaction: S-adenosyl-L-methionine
+ protein L-beta-aspartate = S-adenosyl-L-homocysteine + protein L-beta-aspartate
methyl ester."'
taxonomic_scope:
assessment: MISSING
notes: The taxonomic restrictions in this rule are inconsistent and incoherent.
Condition set 1 (ICMT enzymes via InterPro domains) lacks any taxonomic restriction
and would annotate all domains of life, while condition set 4 (also ICMT enzymes
but via a different FunFam) is restricted to Fungi (NCBITaxon:4751). This creates
an illogical situation where the same enzyme family receives universal annotation
in one condition set but fungi-specific annotation in another. Similarly, condition
set 2 restricts PIMT enzymes to Eukaryota (NCBITaxon:2759) while condition set
5 restricts PIMT enzymes to Metazoa (NCBITaxon:33208), creating overlapping redundant
taxonomic coverage. The fundamental issue is that this multi-family rule conflates
two distinct enzyme families that likely have different phylogenetic distributions.
ICMT enzymes appear to be primarily eukaryotic (involved in CaaX protein processing),
while PIMT enzymes are more broadly distributed across bacteria and eukaryotes.
Condition set 1 is MISSING an appropriate taxonomic restriction for ICMT, which
should likely be restricted to Eukaryota. If the rule were properly split into
separate ICMT and PIMT rules, each could have internally consistent taxonomic
restrictions based on the actual phylogenetic distribution of each enzyme family.
supported_by:
- reference_id: file:rules/arba/ARBA00026372/ARBA00026372.enriched.json
supporting_text: '"Condition set 1: No taxon restriction. Condition set 2: taxon
restriction to Eukaryota (2759). Condition set 4: taxon restriction to Fungi
(4751). Condition set 5: taxon restriction to Metazoa (33208)."'
confidence: 0.95
references:
- id: file:rules/arba/ARBA00026372/ARBA00026372-analysis.yaml
title: Domain overlap analysis for ARBA00026372
findings:
- statement: IPR025770 is a complete subset of IPR007269 with 100% containment
- statement: Average Jaccard similarity across all domain pairs is 0.097
- statement: 9 out of 15 domain pairs show complete disjointness (Jaccard=0.000)
- statement: ICMT-related domains (IPR007269, IPR025770) are completely DISJOINT
from PIMT-related domains (3.40.50.150:FF:000027, 3.40.50.150:FF:000235)
- id: file:rules/arba/ARBA00026372/ARBA00026372-analysis.txt
title: Human-readable domain analysis report
findings:
- statement: Rule contains 5 condition sets with significant structural issues
- statement: Only 1 pair shows >50% overlap (the redundant IPR007269/IPR025770 pair
within CS1)
- id: file:rules/arba/ARBA00026372/ARBA00026372.enriched.json
title: Original ARBA rule definition with enriched annotations
findings:
- statement: Rule annotates to GO:0051998 (protein carboxyl O-methyltransferase
activity)
- statement: Rule created 2021-10-20, last modified 2025-03-21
- statement: Condition sets mix ICMT and PIMT enzyme families with inconsistent
taxonomic restrictions
- id: web:https://pubmed.ncbi.nlm.nih.gov/22195972/
title: Mechanism of isoprenylcysteine carboxyl methylation from the crystal structure
of ICMT
findings:
- statement: ICMT is an integral ER membrane protein with 8 transmembrane domains
- statement: ICMT methylates prenylated cysteine residues at C-terminus of CaaX
proteins
- statement: Methylation increases membrane affinity by neutralizing negative charge
- id: web:https://pubmed.ncbi.nlm.nih.gov/31234761/
title: PIMT-Mediated Protein Repair Mechanism and Implications
findings:
- statement: PIMT is a cytosolic repair enzyme for damaged proteins
- statement: PIMT methylates isoaspartyl residues formed from asparagine deamidation
or aspartate dehydration
- statement: Methylation is transient and initiates repair cycle through succinimide
intermediate
- id: web:https://en.wikipedia.org/wiki/Protein-S-isoprenylcysteine_O-methyltransferase
title: Overview of ICMT structure and function
findings:
- statement: ICMT performs the third step of CaaX protein processing after prenylation
and proteolysis
- statement: Modification is permanent and enhances membrane localization of signaling
proteins
- id: web:https://en.wikipedia.org/wiki/L-isoaspartyl_methyltransferase
title: Overview of PIMT/PCMT1 structure and function
findings:
- statement: PIMT is highly conserved across eukaryotes, archaebacteria, and Gram-negative
bacteria
- statement: Enzyme recognizes and repairs damaged L-isoaspartyl and D-aspartyl
residues in aging proteins