Overview
What this category covers
Everything downstream inherits the quality of the target decision. This category covers the work that comes before design: finding disease-relevant genes in omics and genetic data, establishing whether the resulting protein has a pocket worth pursuing, obtaining a structure when none has been deposited, and placing the protein in the interaction network it operates within. The aim is a target you can defend internally, with the evidence and the gaps both written down.
What we deliver
▸Ranked target candidates with the multi-omics and genetic evidence behind each one
▸Structures for targets that have none, with confidence interpreted region by region
▸Pocket maps with druggability scoring for orthosteric, allosteric and cryptic sites
▸Interaction network context and interface-level tractability assessment
▸A written statement of what the evidence supports and what it does not
Who this is for
▸Discovery teams opening a new indication without a committed target
▸Programmes needing independent scrutiny of a target before major investment
▸Groups holding omics or clinical cohorts they have not fully exploited
▸Projects blocked because the target has no experimental structure
How engagements in this category are run
Target work is scoped around a decision rather than a dataset. We start from the claim you need to be able to make — that this gene drives the disease, that this pocket can be drugged, that this interface can be blocked — and then assemble the evidence that could support or refute it. Public data is used wherever it is adequate, client data is integrated where it sharpens the answer, and each candidate is delivered with the specific experiment that would validate it. Where the available evidence cannot settle the question, that is the finding we report.
Services
5 computational services in this category
A1
Disease Target Mining
Multi-omics integration (genomics, transcriptomics, proteomics) to identify novel disease-relevant targets. GWAS/eQTL analysis, causal gene prioritization, network-based target scoring.
Learn MoreA2
Binding Site & Druggability Assessment
Identification of orthosteric, allosteric, and cryptic pockets. Druggability scoring and hot-spot residue mapping.
Learn MoreA3
AI Protein Structure Prediction
High-accuracy structure prediction for monomers, multimers, and complexes including protein–small molecule, protein–peptide, and protein–nucleic acid.
Learn MoreA4
PPI Network Analysis
Genome-wide PPI network construction, hub protein identification, interaction interface characterization, and druggable PPI target prioritization.
Learn MoreA5
Homology Modeling & Loop Refinement
Template-based 3D model building when experimental structures are unavailable. Multi-template modeling, loop refinement, model validation.
Learn MoreIn Practice
What a project in this category looks like
Three shapes we are asked for most often. Scope is always set after a feasibility review, so timelines below are typical rather than quoted.
Building a first target list for a new indication
The situation
A discovery team opens a programme in a disease area with no committed target and a limited internal dataset.
What we do
Public cohorts are harmonised with any internal data, differential expression and pathway signal are established across the relevant contrasts, genetic causal evidence is added where instruments exist, and every candidate is scored on disease association, specificity, safety exposure and tractability for the intended modality.
What you receive
A ranked list of ten to thirty candidates, each with a two-page dossier, the evidence behind every scoring dimension, and an explicit note of which candidates are weakly supported.
Typical timeline: 4–6 weeks
Testing a target the team already believes in
The situation
A target has internal champions and a proposed mechanism, and the company wants an independent read before committing budget.
What we do
We assemble the evidence for and against: expression in the right tissue and cell type, causal genetic support, dependency data, network position, and whether the protein has a pocket or interface that the intended modality can engage.
What you receive
A written assessment stating what the evidence supports, what it does not, and which single experiment would most efficiently resolve the remaining uncertainty.
Typical timeline: 2–4 weeks
Getting a usable structure for a target with none deposited
The situation
A programme is blocked because the target, or the relevant complex, has no experimental coordinates.
What we do
Complex and monomer prediction with several seeds, confidence interpreted per region and per interface, geometry validation, comparison with homologous structures and any internal biophysical restraints, then preparation for docking or simulation.
What you receive
Ranked models with per-region confidence, a prepared receptor, a pocket assessment, and a clear statement of which parts of the model can carry design decisions.
Typical timeline: 1–3 weeks
Questions
Asked before we scope this work
Can you start from nothing but a disease name?
Yes. Many programmes begin exactly there. Public cohorts and genetic resources carry the first pass; internal data sharpens it. We will tell you during scoping how much difference your own data is likely to make.
How do you avoid simply rediscovering well-known targets?
Known targets are reported and labelled as such rather than hidden, because knowing where your candidate sits relative to the competitive landscape matters. Novelty can be weighted into the scoring if that is the objective.
What does a target dossier actually contain?
Biology and pathway context, disease evidence with the statistics, expression across normal tissue, causal genetic support where available, structural and tractability assessment, known liabilities, and the recommended validation experiment.
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 ↗More
Other service categories
Virtual Screening (Small Molecule)3Small Molecule Design & Optimization2ADMET & Drug-Likeness Prediction4Molecular Dynamics & Simulation3AIDD — Small Molecule Platform2Structure-Based Drug Design (SBDD)3Bioinformatics & Multi-Omics Analysis11Antibody Design & Engineering (AI)5Peptide & Cyclic Peptide Drug Design4Protein Design & Engineering (AI)4ADC & Bioconjugate Design2Compound & Biological Library Design3
Bring us the scientific question, not a service code
Share your target, dataset, sequences or objective. We will recommend a scoped computational plan, the expected deliverables, a timeline and the validation options.
Request a Feasibility Review