The document summarizes the work of the "Novel Enzymatic Activities" group at LABGeM in discovering new enzymatic activities. The group uses several bioinformatics methods like CanOE, ASMC, and reaction molecular signature networks to integrate genomic and structural data to classify enzyme families and predict orphan enzymes. The group has 3 researchers, 2 PhD students, and 1 undergraduate student working on projects like elucidating the function of unknown protein families and exploring archaeal enzyme activities to reduce the number of orphan enzymes and propose new enzymatic functions.

1. Orphan and non-orphan EC number distribution across superkingdoms
Integrated approaches for the discovery of novel enzymatic activities
Guillaume REBOUL, Maria SOROKINA, Jonathan MERCIER, Karine BASTARD, Mark STAM, David VALLENET, Claudine MEDIGUE – CEA, Genoscope, LABGeM
Annotation Rules
Functional annotation rules
UniRule: HAMAP + PIRSF + RuleBase
Pathway Rules
Consistency against biological
processes
Knowledge
Base Rule Engine
Enzyme Activity Discovery workflow
Biological Facts
rule "Missing state"
when
$fact: Fact( present == "no", require == "yes", avoid == "no" )
then
modify( $fact.setState("missing") );
end
rule "require Pathway"
when
$org: Organism()
$path: Pathway(org == $org)
then
Fact fact = new Fact($org, $path);
fact.setRequire("yes");
insert(fact);
end
The “Novel Enzymatic Activities” group
• Group of the LABGeM: Laboratory of Bioinformatics
Analyses for Genomics and Metabolism
• Part of the CEA (French Alternative Energies and
Atomic Energy Commission)
• 3 Researchers
• 2 PhD students
• 1 sandwich placement Master Student
• 1 undergraduate student
• 1 post-doctoral placement available
Pool of
information
able to pull up
Novel
Enzymatic
Activities
Protein or
Domain
Families
Protein
Annotation
and
Sequences
Literature
Enzymatic
Reactions
Own database: NEADB
Summary
?
 iso-functional groups
 multi-functional group
 non-”activity” groups
 Motifs and key residues
for function assignment
3
Promiscuity
& Specificity
Modeling of
compounds
in active sites
+
Full family
 new metabolic
functions and
associated
pathways
4
Metabolic
Role Biochemical
validation
+
Genomic contextRepresentants
enzymes family
1
Define one
reaction
+Multiple alignment
one generic
reaction
A +B <-> A’ + B’
 A family of unknown function
 with experimental evidences
 with one available structure
17 substrats
new
reactions
2
Selection &
Screening
+Enzymatic
screening
Statistical
analysis
+
Family partitioning Potential metabolites
Set of sequences
BLAST PDB
Homology Modeling - MODELLER
3D Models
Cavity Detection - FPOCKET
3D-Active Sites
Structural Alignment - MULTALIGN
Hierarchical Clustering - WEKA
Specificity Determining Residues are determined by a log-likelihood analysis
Pfam unknown
family (DUF 846)
Next Generation Sequencing technology has dramatically increased the number of available sequences in
public databases. At the same time, many enzymatic activities (~22%) are orphans of protein sequence
(Sorokina et al., 2014). The large amount of available protein sequences is an opportunity to discover
enzymes associated to new reactions. We present here an integrated bioinformatics approach to reduce
this lack of knowledge in metabolism and to propose new activity/protein associations for experimental
validation. With this objective, the “New Enzymatic Activity” group of the LABGeM team is developing
several methods. The CanOE method combines genomic and metabolic contexts to predict candidate genes
for orphan enzymes (Smith et al., 2012). Currently, this approach is extended to the detection of
conserved chemical transformation motifs in the metabolism (Sorokina et al., submitted).
From a structural point of view, the ASMC (Active Site Modeling and Clustering) method
finds and compares active site pockets to classify enzymes of a family and detects
important residues for substrate specificity (de Melo-Minardi et al., 2010).
These methods were successful applied to elucidate the enzymatic diversity
of a protein family of unknown function (Bastard et al., 2014). Their
results, associated with present knowledge, must be unified in a
database allowing the elaboration of strategies for the selection of
enzymatic families of interest.
This work is supported by genomic and metabolic network data from
MicroScope, a platform for microbial genome analyses
(Vallenet et al., 2013).
Exploration of archaeal enzyme activities:
ARCHAEOACTOME research project
Literature references
Bastard, K. et al. Revealing the hidden functional diversity of an enzyme family. Nat.
Chem. Biol. 10, 42–9 (2014).
de Melo-Minardi, R. C., Bastard, K. & Artiguenave, F. Identification of subfamily-
specific sites based on active sites modeling and clustering. Bioinformatics 26,
3075–82 (2010).
Smith, A. A. T., Belda, E., Viari, A., Medigue, C. & Vallenet, D. The CanOE strategy:
Integrating genomic and metabolic contexts across multiple prokaryote genomes to
find candidate genes for orphan enzymes. PLoS Comput. Biol. 8, (2012).
Sorokina, M., Stam, M., Médigue, C., Lespinet, O. & Vallenet, D. Profiling the orphan
enzymes. Biol. Direct 9, 10 (2014).
Vallenet, D. et al. MicroScope--an integrated microbial resource for the curation and
comparative analysis of genomic and metabolic data. Nucleic Acids Res. 41, D636–
47 (2013).
Sorokina et al. A novel metabolic network representation for the discovery of
conserved modules of chemical transformations. Submitted.
Mercier, J., Vallenet, D. GROOLS: Reactive Graph Reasoning for Genome
Annotation. RuleML 2015 Conference
Active Sites Classification (ASMC)
The CanOE strategy
Reaction Molecular Signature Network
The dynamics of enzyme discoveryMicroScope
From genomes to biological
systems
Reactions sharing a same RMS
Reaction Network reduction
into a RMS Network
Microbial genome
analysis Metabolic network
>3,900 genomes
1-10 Mb
ASMC method
Classification of a family
into groups of similar active sites
NEA team
Workbench
Data
Integration
Orphan
Enzymes
Grools
Structural
Analysis
Metabolic
Network
MicroScope