Domine

Specification

• CAS No.: 94823-32-2
• Molecular Formula: C17H38N2
• Molecular Weight: 270.5 g/mol
• IUPAC Name: N'-dodecyl-N,N,N'-trimethylethane-1,2-diamine
• Standard InChI: InChI=1S/C17H38N2/c1-5-6-7-8-9-10-11-12-13-14-15-19(4)17-16-18(2)3/h5-17H2,1-4H3
• Standard InChI Key: LLJOASAHEBQANS-UHFFFAOYSA-N
• Canonical SMILES: CCCCCCCCCCCCN(C)CCN(C)C

Introduction

Overview of DOMINE’s Structure and Scope

DOMINE is structured to consolidate both experimentally validated and computationally predicted DDIs. Its primary objective is to bridge the gap between theoretical predictions and empirical validation, offering a platform for cross-referencing interactions across diverse biological contexts. The database categorizes interactions into two broad classes:

Known DDIs

Known DDIs are derived from high-resolution 3D structural data, primarily sourced from iPfam and 3did databases. These interactions are experimentally confirmed through techniques such as X-ray crystallography and nuclear magnetic resonance (NMR) spectroscopy. As of the latest update, DOMINE contains 6,634 known DDIs, reflecting a 35% increase from earlier versions due to advancements in structural biology.

Predicted DDIs

Predicted DDIs are inferred using computational algorithms that analyze sequence homology, phylogenetic profiles, and gene ontology (GO) term co-occurrence. Notable methods integrated into DOMINE include:

• GPE (Genome Phylogenetic Evolution): Leverages evolutionary conservation patterns.

• DIPD (Domain Interaction Prediction using Deep learning): Employs neural networks to predict interactions.

• K-GIDDI (Knowledge-Guided Inference of Domain-Domain Interactions): Integrates prior biological knowledge with machine learning.

These predictions account for 21,620 entries in DOMINE, classified into high-, medium-, and low-confidence tiers based on methodological robustness and cross-validation.

Data Integration and Classification Framework

Data Sources

DOMINE aggregates data from 15 repositories and prediction tools (Table 1), ensuring breadth and depth in its coverage. Key sources include:

Source Type	Examples	Contribution to DOMINE
Experimental	iPfam, 3did	6,634 known DDIs
Computational	GPE, DIPD, K-GIDDI	21,620 predicted DDIs
Hybrid	DomainGA, Insite	Integrates structural & evolutionary data

Table 1: Primary data sources contributing to DOMINE’s DDI repository.

Confidence Classification Schema

DOMINE employs a tripartite confidence system to evaluate predicted DDIs:

• High-Confidence Predictions (HCP): DDIs predicted by ≥2 independent methods or supported by structural/functional evidence.

• Medium-Confidence Predictions (MCP): DDIs with GO term co-occurrence but lacking multi-method validation.

• Low-Confidence Predictions (LCP): DDIs inferred from single-method predictions without ancillary evidence.

This stratification aids researchers in prioritizing interactions for experimental validation, reducing resource allocation to spurious associations.

Applications in Biomedical Research

Protein Interaction Network (PIN) Prediction

DOMINE’s DDIs underpin the construction of PINs, which map interactomes for organisms lacking comprehensive experimental data. For instance, DDIs from DOMINE have been used to predict PINs in Plasmodium falciparum, identifying novel drug targets for malaria.

Drug Discovery and Target Identification

By elucidating interaction interfaces, DOMINE facilitates the design of small-molecule inhibitors. A case study involving the EGFR kinase domain revealed cryptic binding sites disruptable by allosteric inhibitors, a strategy now employed in oncology therapeutics.

Functional Annotation of Unknown Proteins

DOMINE enables the functional annotation of uncharacterized proteins via "guilt-by-association." For example, the yeast protein YGR234W was annotated as a RNA helicase after its DDI with known helicase domains was cataloged in DOMINE.

Limitations and Future Directions

Despite its utility, DOMINE faces challenges:

Planned enhancements include:

• Integration of multi-domain classification systems.

• Real-time updating via automated literature mining tools.