Ref A4
Computational Service
Confidence Reported
Scoped After Feasibility Review
Service Overview
Protein–protein interaction analysis serves two different purposes and we treat them separately. At the network level, it places a protein in its functional context: what it connects to, which modules it sits in, and whether disrupting it would propagate usefully or catastrophically. At the structural level, it characterises a specific interface — its area, hot-spot residues and sub-pockets — to judge whether that interaction can be blocked by a small molecule, a peptide or a biologic. Most programmes need both answers.
Key Features & Advantages
Network and Interface in One Study
Topological analysis tells you which interactions matter biologically; structural modelling tells you which of those can realistically be drugged. Neither answer is useful alone.
Confidence-Filtered Interactomes
Interactions are filtered by evidence type and confidence rather than pooled indiscriminately, so hubs are not created by text-mining artefacts.
Hot-Spot Driven Tractability
Interfaces are scored on hot-spot clustering and pocket presence — the features that separate the minority of PPIs with small-molecule potential from the rest.
Perturbation Consequences Modelled
We report what else is predicted to be affected when the interaction is broken, which is often where the toxicity risk hides.
Where This Service Is Used
1Judging whether a specific protein–protein interaction is a viable drug target
2Placing a novel or poorly characterised protein into a functional pathway context
3Finding the most connected, disease-relevant node in a pathway for intervention
4Designing peptide or biologic inhibitors against a defined interface
5Assessing which downstream complexes a candidate degrader or inhibitor would disturb
Methods and Techniques Applied
1Interactome assembly from curated databases with evidence-type filtering and confidence thresholds
2Topological analysis: degree, betweenness, clustering, module detection and disease-gene enrichment within modules
3Structural modelling of selected interfaces using multimer prediction and protein–protein docking
4Interface characterisation: buried surface area, shape complementarity, hot-spot residues, sub-pocket detection
5Alanine-scanning style energetic decomposition to rank residues by contribution to binding
6Tractability call per interface, with the modality most likely to succeed stated explicitly
What We Need From You
1The protein or pathway of interest, with species and isoform specified
2Known partners, affinities or mutational data that constrain the interface
3Structures of the partners if available; otherwise we predict them as part of the work
4Whether you intend a small molecule, peptide, biologic or degrader, since tractability criteria differ
What You Receive
1Interaction network files with confidence annotation, in standard network exchange formats
2Ranked partner table with network metrics, disease relevance and evidence type
3Structural models of the prioritised interfaces with hot-spot residues annotated
4Tractability assessment per interface with the recommended modality and the reasoning
5Figures and session files suitable for internal presentation and grant or investor material
What We Compute and Report
1Interaction confidence score per protein pair, with the evidence type behind each
2Network metrics per node: degree, betweenness, clustering coefficient and module assignment
3Buried surface area, shape complementarity and contact counts for each modelled interface
4Per-residue binding energy decomposition ranking hot-spot residues at the interface
5Predicted consequence of disrupting the interaction across the connected sub-network
Questions We Are Asked Before Scoping
Are predicted interactions reliable enough to act on?
▸Curated, experimentally supported interactions are reasonably reliable; purely predicted or text-mined ones are not, and are labelled accordingly. Acting on a single predicted interaction without validation is not advisable.
Can most PPIs be drugged with small molecules?
▸No. Most interfaces are large, flat and lack a concentrated hot spot. Our tractability assessment exists precisely to identify the minority that are viable, and to say so when yours is not.
How accurate is complex prediction for a novel interaction?
▸Accuracy varies widely with interface conservation and coevolutionary signal. Interface confidence scores are reported per model, and weak predictions are flagged as hypotheses requiring crosslinking or mutagenesis support.
Interpretation and Limitations
Computational predictions prioritise hypotheses; they do not independently establish binding, biological activity, therapeutic efficacy, safety or clinical validity. Performance depends on input quality, target class, the structural evidence available and the applicability domain of the selected methods. Scores used for ranking are comparative devices and should not be read as measured quantities. All deliverables are for research use only and are not intended for diagnostic or therapeutic use.
Is This the Right Service for You?
Choose this when
You need to know what a protein connects to before deciding to perturb it
A specific interaction is the proposed intervention point and its tractability is unknown
Off-target or downstream consequences of disrupting an interaction need to be anticipated
Consider another service when
The interface is established and you need a molecule to block it
De Novo Small Molecule Design A peptide or biologic is the intended modality for the interface
Linear Therapeutic Peptide Design What Goes Wrong, and How We Handle It
These are the failure modes that produce confident-looking results which do not survive the next experiment. Each one is designed against rather than discovered afterwards.
Text-mined interactions inflating a protein into an apparent hub
Edges are filtered by evidence type and confidence before any topology metric is computed.
Assuming a predicted complex is real because the model looks plausible
Interface confidence is reported separately, and weak predictions are delivered as hypotheses for crosslinking or mutagenesis.
Treating every interface as druggable
Hot-spot concentration and sub-pocket presence are quantified; interfaces without them are reported as unsuitable for small molecules.
A Worked Example
Triaging a pathway for an intervention point
The situation
A team had a validated pathway but no view on which node to target or with what modality.
What the work involved
A confidence-filtered interactome was built around the pathway, modules and bottlenecks were identified, and the six most disease-relevant interfaces were modelled and characterised for hot spots and sub-pockets.
What came back
One interface with a concentrated hot spot and a small-molecule-sized pocket, three suitable only for peptides or biologics, and two rejected as flat and featureless.
Illustrative of the shape this work takes. Details are generalised, no client or target is identified, and scope and outcome vary with the system in question.
What Published Work Reports
The results below were published by independent groups and are cited to show what these computational approaches have achieved in the literature. They are not ICDMO project results, and outcomes vary substantially between targets, datasets and programmes. We do not promise comparable results for any particular project.
Around three quarters of predicted ligand poses fall within 2 Å
Current deep-learning structure prediction, evaluated on an independent protein–ligand benchmark, placed roughly 76% of ligands within 2 Å of the experimental pose, and reported approximately 80% by coordinate error — a substantial improvement over classical docking baselines on the same set, achieved without supplying a binding site. The corollary matters as much as the headline: roughly a quarter of cases still fall outside that threshold, which is why we report confidence per interface rather than delivering a single model.
Abramson et al., Nature, 2024 ↗Service Workflow
1
Scientific Consultation
We discuss the objective, the decision it supports and the data you hold. Nothing is quoted before we understand what the result has to answer.
2
Feasibility Review
We check that the data, target class and available structural evidence can support the proposed approach, and say plainly where they cannot.
3
Technical Proposal
Scope, methods, deliverables, timeline and price are confirmed in writing, together with the limitations we expect to report.
4
Computation and QC
The analysis runs on pinned tool versions in a containerised environment, with quality checkpoints and method comparison where that is informative.
5
Report and Follow-Up
You receive the ranked results, the documented methods, the limitations and a proposed validation plan, plus support in interpreting them.
Service Process
1
Online Consultation
›
2
Solution Matching
›
3
Service Contract
›
4
Project Execution
›
5
Data Delivery
Specifications & Turnaround
Note: ICDMO delivers computational (in silico) services. Results prioritise hypotheses for testing; they do not by themselves establish binding, biological activity, efficacy or safety. All deliverables are for research use only and are not intended for diagnostic or therapeutic use. Downstream experimental work is scoped separately and executed through qualified partners.
Get a Custom Quote
Our scientific team responds within 24 hours with a detailed technical proposal and pricing tailored to your project.
Contact Us Online ConsultationStandard Deliverables
✓Sequence-verified final product
✓Full QC documentation
✓Certificate of Analysis (CoA)
✓Electronic data files
✓Project completion report
Why ICDMO?
Computational CDMO Focus
Reproducible, Versioned Pipelines
Confidence & Limitations Reported
Experiment-Aware Recommendations
Response Within One Business Day