ANTI-INFECTIVE DISCOVERY · CATALOG 2026.09

Antiviral & Antibacterial
Drug Discovery

Advance anti-infective discovery with molecular activity prediction, species-specific screening and study design. Connect target and assay context to candidate prioritization and focused experimental validation.

Explore research solutions Plan your study

From candidate triage to custom modeling, focus your next round of experiments.

03Research areas

10Specialist capabilities

26Research solutions

12Activity & developability endpoints

01 / RESEARCH EXPERTISE

Expertise across anti-infective discovery

Integrate target activity, selectivity, infection dynamics and experimental validation across antiviral, antibacterial and lead-optimization programs.

02 / RESEARCH SOLUTIONS

Find the service for your next study

Select by target, species or development stage, with clearly defined data requirements, methods and deliverables.

AV-TGT-001

Antiviral discovery / Target activity

Activity / property prediction

NA inhibitor prioritization

Rank NA-directed small molecules for biochemical confirmation.

Inputs

Compound IDs and SMILES; NA target context; optional biochemical measurements.

Deliverables / deliverables

Predicted pIC50 / IC50 (nM), domain flags, nearest-neighbor evidence and ranked candidates.

Methods & service scope

Model / method: ECFP4 plus q-RASAR similarity features; 300-tree random forest; three-fingerprint applicability assessment.
Scope & considerations: Supports candidate ranking within influenza NA-related chemical space. Cellular efficacy and other viral targets require separate calibration and validation.

n 1,143ρ 0.809RMSE 1.070

Development benchmark: scaffold cross-validation · Historical model-development results; project applicability is assessed through independent validation.

Add to project

AV-TGT-002

Antiviral discovery / Target activity

Activity / property prediction

PA endonuclease inhibitor prioritization

Prioritize compounds and related chemical series for PA-targeted programs.

Inputs

Compound IDs, SMILES and PA target context.

Deliverables / deliverables

Predicted pIC50 / IC50 (nM), ranking, domain flags and analog evidence.

Methods & service scope

Model / method: ECFP4 plus q-RASAR; ridge regression; three-fingerprint domain assessment.
Scope & considerations: Developed on 252 compounds. New scaffolds and highly potent candidates require analog review and biochemical confirmation.

n 252ρ 0.734RMSE 0.728

Development benchmark: scaffold cross-validation · Historical model-development results; project applicability is assessed through independent validation.

Add to project

AV-DYN-001

Antiviral discovery / Infection dynamics

Exploratory simulation

Viral-load and treatment-timing simulation

Compare exposure and treatment timing to support sampling design.

Inputs

Measured or explicitly assumed EC50, concentration, half-life, start time and system-specific kinetic parameters.

Deliverables / deliverables

Simulated viral-load trajectories, peak changes and scenario comparisons; AUC requires a defined observation window.

Methods & service scope

Model / method: Four-compartment target-cell ODE with eclipse phase, Hill inhibition and constant or decaying exposure.
Scope & considerations: Uses H3N2 literature parameters as a starting point. Calibrate kinetics for the intended cell system and link CPE to matched experimental readouts.

Add to project

AV-SCR-001

Antiviral discovery / Screening & assay context

Precomputed case

Docking and screening enrichment review

Review QSAR–docking ranking agreement to allocate follow-up screening capacity.

Inputs

Embedded library, target and selection thresholds; a client library requires a separate computation.

Deliverables / deliverables

Ranking plots, candidate tables, docking coverage and enrichment curves against proxy endpoints.

Methods & service scope

Model / method: Precomputed QSAR and Vina results for a 500-compound NA/PA library; threshold and enrichment analysis.
Scope & considerations: The reference case covers 500 molecules with docking for approximately 100 prioritized candidates. Client libraries require a dedicated calculation; enrichment is measured against computational proxy endpoints.

Add to project

AV-SCR-002

Antiviral discovery / Screening & assay context

System & data assessment

Antiviral cell-system data passport

Assess cell-system data coverage and identify measurements needed for modeling.

Inputs

Cell line, virus, assay endpoint and existing client data.

Deliverables / deliverables

System passport, data-coverage table, supported tasks and gap list.

Methods & service scope

Model / method: Curation of cell-line CC50 coverage, use cases and omics context.
Scope & considerations: Supports data-coverage assessment and assay-system selection. Infection-phenotype and cross-cell toxicity predictions require matched data.

Add to project

AV-SCR-003

Antiviral discovery / Screening & assay context

Precomputed case

NA500 screening case explorer

Demonstrate candidate review spanning molecular activity, structure and kinetics.

Inputs

Embedded historical library; domain, docking-coverage and ranking filters.

Deliverables / deliverables

Multi-metric plots, top-candidate tables, compound evidence cards and historical traceability.

Methods & service scope

Model / method: Historical 500-compound model snapshot with a 101-compound docking subset.
Scope & considerations: Provides traceable results from a historical computational case. Infection dynamics are simulated; new candidates require assessment with the agreed model version.

Add to project

AB-MIC-001

Antibacterial discovery / Species-specific MIC

Activity / property prediction

E. coli MIC prioritization

Rank candidate antibacterial activity against E. coli for single-species or panel screening.

Inputs

Compound IDs, SMILES and target species; molecular weight for mass-concentration conversion.

Deliverables / deliverables

Predicted pMIC, MIC (µM / µg/mL), domain flags and within-species rankings.

Methods & service scope

Model / method: Exact MIC records from ChEMBL; ECFP4-2048 random forest; scaffold CV and nearest-neighbor domain assessment.
Scope & considerations: Supports species-level candidate prioritization. Strain-specific susceptibility categories, bactericidal concentrations and time-kill effects require experimental confirmation.

n 31,676ρ 0.719RMSE 0.720

Development benchmark: scaffold cross-validation · Historical model-development results; project applicability is assessed through independent validation.

Add to project

AB-MIC-002

Antibacterial discovery / Species-specific MIC

Activity / property prediction

K. pneumoniae MIC prioritization

Rank candidate antibacterial activity against K. pneumoniae for single-species or panel screening.

Inputs

Compound IDs, SMILES and target species; molecular weight for mass-concentration conversion.

Deliverables / deliverables

Predicted pMIC, MIC (µM / µg/mL), domain flags and within-species rankings.

Methods & service scope

Model / method: Exact MIC records from ChEMBL; ECFP4-2048 random forest; scaffold CV and nearest-neighbor domain assessment.
Scope & considerations: Supports species-level candidate prioritization. Strain-specific susceptibility categories, bactericidal concentrations and time-kill effects require experimental confirmation.

n 12,015ρ 0.754RMSE 0.701

Development benchmark: scaffold cross-validation · Historical model-development results; project applicability is assessed through independent validation.

Add to project

AB-MIC-003

Antibacterial discovery / Species-specific MIC

Activity / property prediction

P. aeruginosa MIC prioritization

Rank candidate antibacterial activity against P. aeruginosa for single-species or panel screening.

Inputs

Compound IDs, SMILES and target species; molecular weight for mass-concentration conversion.

Deliverables / deliverables

Predicted pMIC, MIC (µM / µg/mL), domain flags and within-species rankings.

Methods & service scope

Model / method: Exact MIC records from ChEMBL; ECFP4-2048 random forest; scaffold CV and nearest-neighbor domain assessment.
Scope & considerations: Supports species-level candidate prioritization. Strain-specific susceptibility categories, bactericidal concentrations and time-kill effects require experimental confirmation.

n 20,022ρ 0.698RMSE 0.676

Development benchmark: scaffold cross-validation · Historical model-development results; project applicability is assessed through independent validation.

Add to project

AB-MIC-004

Antibacterial discovery / Species-specific MIC

Activity / property prediction

S. aureus MIC prioritization

Rank candidate antibacterial activity against S. aureus for single-species or panel screening.

Inputs

Compound IDs, SMILES and target species; molecular weight for mass-concentration conversion.

Deliverables / deliverables

Predicted pMIC, MIC (µM / µg/mL), domain flags and within-species rankings.

Methods & service scope

Model / method: Exact MIC records from ChEMBL; ECFP4-2048 random forest; scaffold CV and nearest-neighbor domain assessment.
Scope & considerations: Supports species-level candidate prioritization. Strain-specific susceptibility categories, bactericidal concentrations and time-kill effects require experimental confirmation.

n 48,610ρ 0.714RMSE 0.731

Development benchmark: scaffold cross-validation · Historical model-development results; project applicability is assessed through independent validation.

Add to project

AB-AMR-001

Antibacterial discovery / Resistance markers

Mechanistic / rule-based analysis

Acquired-resistance marker screen

Screen assembled sequences for known resistance markers to guide full annotation.

Inputs

Assembled contigs/genome FASTA or nucleotide sequence; k-mer length.

Deliverables / deliverables

Markers, associated drug classes, k-mer containment, partial hits and review flags.

Methods & service scope

Model / method: Canonical k-mer containment against 42 selected AMRFinderPlus markers.
Scope & considerations: Covers reference fragments from 42 acquired-resistance markers. Alleles, point mutations and susceptibility phenotypes require further confirmation; a negative screen does not imply susceptibility.

Add to project

AB-DYN-001

Antibacterial discovery / Antibacterial dynamics

Exploratory simulation

Antibacterial exposure and resistance simulation

Compare dosing interval, relative exposure and resistance assumptions on bacterial burden.

Inputs

Exposure and MIC in consistent units, half-life, interval, resistance fold change, mutation and growth parameters.

Deliverables / deliverables

Simulated bacterial burden, resistant fraction, PK/PD indices and resistance-dominance timing maps.

Methods & service scope

Model / method: One-compartment oral PK with susceptible/resistant population growth–kill ODEs.
Scope & considerations: Designed for research scenario comparisons. Parameters require calibration to strain-specific exposure and killing data; clinical dosing is outside scope.

Add to project

DS-SAF-001

Shared screening & optimization / Selectivity & developability

Activity / property prediction

Pooled cytotoxicity and selectivity reference

Flag potentially nonselective compounds and prioritize cytotoxicity confirmation.

Inputs

SMILES; optional paired EC50/CC50 measurements in compatible units and assay context.

Deliverables / deliverables

Predicted pCC50 / CC50, domain flags; exploratory selectivity proxies or measured paired SI.

Methods & service scope

Model / method: Pooled pCC50 random forest plus q-RASAR; joint review with activity estimates.
Scope & considerations: A pooled estimate is not target-cell CC50. CC50/biochemical IC50 is a proxy, not a validated antiviral SI.

n 33,641ρ 0.704RMSE 0.476

Development benchmark: scaffold cross-validation · Historical model-development results; project applicability is assessed through independent validation.

Add to project

DS-SAF-002

Shared screening & optimization / Selectivity & developability

Activity / property prediction

Aqueous-solubility triage

Add aqueous-solubility triage to antibacterial and antiviral lead prioritization.

Inputs

Compound IDs and SMILES.

Deliverables / deliverables

Predicted logS (log10(mol/L)), domain flags, risk ranking and confirmation list.

Methods & service scope

Model / method: AqSolDB; ECFP4-2048 random forest, 300-tree parent / 100-tree page subset.
Scope & considerations: pH, solid form and assay conditions can alter solubility; the large error supports early triage.

n 12,930ρ 0.599RMSE 1.898

Development benchmark: scaffold cross-validation · Historical model-development results; project applicability is assessed through independent validation.

Add to project

DS-SAF-003

Shared screening & optimization / Selectivity & developability

Activity / property prediction

hERG inhibition triage

Add herg inhibition triage to antibacterial and antiviral lead prioritization.

Inputs

Compound IDs and SMILES.

Deliverables / deliverables

Predicted pIC50 (−log10(mol/L)), domain flags, risk ranking and confirmation list.

Methods & service scope

Model / method: ChEMBL / KCNH2; ECFP4-2048 random forest, 300-tree parent / 100-tree page subset.
Scope & considerations: Predicts in-vitro inhibition tendency, not clinical cardiac safety.

n 9,716ρ 0.606RMSE 0.704

Development benchmark: scaffold cross-validation · Historical model-development results; project applicability is assessed through independent validation.

Add to project

DS-SAF-004

Shared screening & optimization / Selectivity & developability

Activity / property prediction

CYP3A4 inhibition triage

Add cyp3a4 inhibition triage to antibacterial and antiviral lead prioritization.

Inputs

Compound IDs and SMILES.

Deliverables / deliverables

Predicted pIC50 (−log10(mol/L)), domain flags, risk ranking and confirmation list.

Methods & service scope

Model / method: ChEMBL / CYP3A4; ECFP4-2048 random forest, 300-tree parent / 100-tree page subset.
Scope & considerations: Does not cover induction, time-dependent inhibition or human drug interactions.

n 5,490ρ 0.573RMSE 0.652

Development benchmark: scaffold cross-validation · Historical model-development results; project applicability is assessed through independent validation.

Add to project

DS-SAF-005

Shared screening & optimization / Selectivity & developability

Activity / property prediction

CYP2D6 inhibition triage

Add cyp2d6 inhibition triage to antibacterial and antiviral lead prioritization.

Inputs

Compound IDs and SMILES.

Deliverables / deliverables

Predicted pIC50 (−log10(mol/L)), domain flags, risk ranking and confirmation list.

Methods & service scope

Model / method: ChEMBL / CYP2D6; ECFP4-2048 random forest, 300-tree parent / 100-tree page subset.
Scope & considerations: Does not establish genotype-specific metabolism or clinical drug interactions.

n 2,993ρ 0.579RMSE 0.695

Development benchmark: scaffold cross-validation · Historical model-development results; project applicability is assessed through independent validation.

Add to project

DS-SAF-006

Shared screening & optimization / Selectivity & developability

Activity / property prediction

Human liver microsomal stability triage

Add human liver microsomal stability triage to antibacterial and antiviral lead prioritization.

Inputs

Compound IDs and SMILES.

Deliverables / deliverables

Predicted log half-life (log10(min)), domain flags, risk ranking and confirmation list.

Methods & service scope

Model / method: TDC / Obach; ECFP4-2048 random forest, 300-tree parent / 100-tree page subset.
Scope & considerations: Only 601 training compounds; high uncertainty, with no direct conversion to human half-life.

n 601ρ 0.449RMSE 0.488

Development benchmark: scaffold cross-validation · Historical model-development results; project applicability is assessed through independent validation.

Add to project

DS-DEC-001

Shared screening & optimization / Decision support & experiment design

Mechanistic / rule-based analysis

Multi-model anti-infective triage

Combine activity, toxicity and domain evidence into reviewable testing priorities.

Inputs

Compound list, antiviral/antibacterial branch, activity/window gates and target species or protein.

Deliverables / deliverables

Prioritize / review / defer lists, endpoint estimates, domain flags, reasons and CSV.

Methods & service scope

Model / method: Seven endpoints: NA, PA, CC50 and four MIC models; rule-based funnel with decision reasons.
Scope & considerations: Uses biochemical activity or species-specific MIC for triage, with separate target and species assessments. Selectivity is a proxy; out-of-domain candidates require dedicated validation.

Add to project

DS-DEC-002

Shared screening & optimization / Decision support & experiment design

Combination-effect analysis

Combination reference surfaces and synergy analysis

Construct no-interaction references and analyze measured combination responses.

Inputs

Single-agent IC50 and Hill slopes, concentration grid, measured combination inhibition; MIC checkerboards require MIC data.

Deliverables / deliverables

8×8 expected inhibition surfaces, observed deviations and equivalent-dose indices, interpreted by endpoint.

Methods & service scope

Model / method: Single-agent Hill curves; Loewe, Bliss and HSA reference surfaces; equivalent-dose analysis.
Scope & considerations: Actual synergy requires combination measurements. Equivalent-dose indices are not automatically MIC-based FICI; thresholds must match the assay.

Add to project

DS-DEC-003

Shared screening & optimization / Decision support & experiment design

Mechanistic / rule-based analysis

Next-batch experiment selection

Allocate limited testing capacity across activity, exploration and diversity.

Inputs

Embedded 500-compound antiviral or 1,000-compound antibacterial pool, weights and batch size; client pools need features and predictions.

Deliverables / deliverables

Compound batch, selection rationale, activity–uncertainty plot and CSV.

Methods & service scope

Model / method: Weighted greedy selection: exploitation, forest-variance proxy and structural diversity; embedded precomputed pools.
Scope & considerations: Tree variance guides exploration. Model updates and adaptation to client libraries are scoped separately. Deliverables prioritize experiments rather than guarantee activity.

Add to project

DS-DEC-004

Shared screening & optimization / Decision support & experiment design

Exploratory simulation

Antiviral exposure and resistance simulation

Explore how exposure and dosing interval affect wild-type/resistant viral competition.

Inputs

EC50, relative exposure, half-life, interval, mutation parameters and resistance fold change.

Deliverables / deliverables

Viral load and resistant-fraction trajectories, 72-scenario scan and resistance-dominance timing.

Methods & service scope

Model / method: One-compartment PK with two-strain target-cell-limited ODEs and resistance/mutation assumptions.
Scope & considerations: Compares exposure and resistance-selection scenarios, with calibration to matched system data. Deliverables are not clinical doses or real-world resistance probabilities.

Add to project

DS-MOA-001

Shared screening & optimization / Host-mechanism support

Mechanistic / rule-based analysis

Host-protein modification motif scan

Nominate phosphorylation and acetylation-context sites for host-mechanism studies.

Inputs

Protein FASTA or amino-acid sequence.

Deliverables / deliverables

Candidate sites, sequence context, motif scores and site map.

Methods & service scope

Model / method: Hand-built PKA, PKC, CDK, CK2 and ATM motif matrices plus lysine-context scoring.
Scope & considerations: Not a trained model; scores are not modification probabilities and omit accessibility, localization and experimental confirmation.

Add to project

DS-CUS-001

Shared screening & optimization / Custom modeling & validation

Custom development · feasibility assessment

Biochemical-to-cellular efficacy calibration

Build and test an IC50-to-EC50 bridge for a defined target/cell system, then evaluate CPE or viral-load outputs.

Inputs

Structures, paired IC50/EC50, CC50, cell line and assay context; optional permeability/protein-binding data.

Deliverables / deliverables

Calibrated model, error/domain report and cellular estimates only after prespecified gates are met.

Methods & service scope

Model / method: Matched biochemical/cellular data, regression baselines, exposure corrections and independent holdout validation.
Scope & considerations: Requires paired biochemical and cellular measurements. Feasibility assessment precedes development of a system-specific calibration model.

Add to project

DS-CUS-002

Shared screening & optimization / Custom modeling & validation

Custom development · feasibility assessment

New-target and new-species model development

Extend the curation, QSAR and domain-assessment workflow to a client target or species.

Inputs

Target, endpoint, authorized structure–activity data and an independent test set.

Deliverables / deliverables

Feasibility decision, model card, training/validation versions, prediction-interface specification or stop recommendation.

Methods & service scope

Model / method: Data assessment, endpoint harmonization, deduplication, scaffold/study holdouts, model comparison and applicability calibration.
Scope & considerations: The standard MIC panel covers four species. Expansion is evaluated against data coverage, endpoint consistency and independent validation.

Add to project

DS-CUS-003

Shared screening & optimization / Custom modeling & validation

Custom development · feasibility assessment

Independent validation and assay feedback

Test ranking utility on unseen data before scaling screening or updating models.

Inputs

Frozen model, training provenance and membership, external or prospective assay data, budget and prespecified metrics.

Deliverables / deliverables

Independent performance report, enrichment/recall, error analysis, out-of-domain behavior and go/revise recommendation.

Methods & service scope

Model / method: Frozen model versions, compound/scaffold/study overlap checks, blinded or external evaluation, baseline comparisons and assay-feedback analysis.
Scope & considerations: Offered as a scoped validation engagement. The protocol, data independence and success criteria are agreed before analysis.

Add to project

Research area	Capability	CAT / Solution	Analysis type	Inputs	Deliverables
Antiviral discovery	Target activity	`AV-TGT-001` NA inhibitor prioritization	Activity / property prediction	Compound IDs and SMILES; NA target context; optional biochemical measurements.	Predicted pIC50 / IC50 (nM), domain flags, nearest-neighbor evidence and ranked candidates.
Antiviral discovery	Target activity	`AV-TGT-002` PA endonuclease inhibitor prioritization	Activity / property prediction	Compound IDs, SMILES and PA target context.	Predicted pIC50 / IC50 (nM), ranking, domain flags and analog evidence.
Antiviral discovery	Infection dynamics	`AV-DYN-001` Viral-load and treatment-timing simulation	Exploratory simulation	Measured or explicitly assumed EC50, concentration, half-life, start time and system-specific kinetic parameters.	Simulated viral-load trajectories, peak changes and scenario comparisons; AUC requires a defined observation window.
Antiviral discovery	Screening & assay context	`AV-SCR-001` Docking and screening enrichment review	Precomputed case	Embedded library, target and selection thresholds; a client library requires a separate computation.	Ranking plots, candidate tables, docking coverage and enrichment curves against proxy endpoints.
Antiviral discovery	Screening & assay context	`AV-SCR-002` Antiviral cell-system data passport	System & data assessment	Cell line, virus, assay endpoint and existing client data.	System passport, data-coverage table, supported tasks and gap list.
Antiviral discovery	Screening & assay context	`AV-SCR-003` NA500 screening case explorer	Precomputed case	Embedded historical library; domain, docking-coverage and ranking filters.	Multi-metric plots, top-candidate tables, compound evidence cards and historical traceability.
Antibacterial discovery	Species-specific MIC	`AB-MIC-001` E. coli MIC prioritization	Activity / property prediction	Compound IDs, SMILES and target species; molecular weight for mass-concentration conversion.	Predicted pMIC, MIC (µM / µg/mL), domain flags and within-species rankings.
Antibacterial discovery	Species-specific MIC	`AB-MIC-002` K. pneumoniae MIC prioritization	Activity / property prediction	Compound IDs, SMILES and target species; molecular weight for mass-concentration conversion.	Predicted pMIC, MIC (µM / µg/mL), domain flags and within-species rankings.
Antibacterial discovery	Species-specific MIC	`AB-MIC-003` P. aeruginosa MIC prioritization	Activity / property prediction	Compound IDs, SMILES and target species; molecular weight for mass-concentration conversion.	Predicted pMIC, MIC (µM / µg/mL), domain flags and within-species rankings.
Antibacterial discovery	Species-specific MIC	`AB-MIC-004` S. aureus MIC prioritization	Activity / property prediction	Compound IDs, SMILES and target species; molecular weight for mass-concentration conversion.	Predicted pMIC, MIC (µM / µg/mL), domain flags and within-species rankings.
Antibacterial discovery	Resistance markers	`AB-AMR-001` Acquired-resistance marker screen	Mechanistic / rule-based analysis	Assembled contigs/genome FASTA or nucleotide sequence; k-mer length.	Markers, associated drug classes, k-mer containment, partial hits and review flags.
Antibacterial discovery	Antibacterial dynamics	`AB-DYN-001` Antibacterial exposure and resistance simulation	Exploratory simulation	Exposure and MIC in consistent units, half-life, interval, resistance fold change, mutation and growth parameters.	Simulated bacterial burden, resistant fraction, PK/PD indices and resistance-dominance timing maps.
Shared screening & optimization	Selectivity & developability	`DS-SAF-001` Pooled cytotoxicity and selectivity reference	Activity / property prediction	SMILES; optional paired EC50/CC50 measurements in compatible units and assay context.	Predicted pCC50 / CC50, domain flags; exploratory selectivity proxies or measured paired SI.
Shared screening & optimization	Selectivity & developability	`DS-SAF-002` Aqueous-solubility triage	Activity / property prediction	Compound IDs and SMILES.	Predicted logS (log10(mol/L)), domain flags, risk ranking and confirmation list.
Shared screening & optimization	Selectivity & developability	`DS-SAF-003` hERG inhibition triage	Activity / property prediction	Compound IDs and SMILES.	Predicted pIC50 (−log10(mol/L)), domain flags, risk ranking and confirmation list.
Shared screening & optimization	Selectivity & developability	`DS-SAF-004` CYP3A4 inhibition triage	Activity / property prediction	Compound IDs and SMILES.	Predicted pIC50 (−log10(mol/L)), domain flags, risk ranking and confirmation list.
Shared screening & optimization	Selectivity & developability	`DS-SAF-005` CYP2D6 inhibition triage	Activity / property prediction	Compound IDs and SMILES.	Predicted pIC50 (−log10(mol/L)), domain flags, risk ranking and confirmation list.
Shared screening & optimization	Selectivity & developability	`DS-SAF-006` Human liver microsomal stability triage	Activity / property prediction	Compound IDs and SMILES.	Predicted log half-life (log10(min)), domain flags, risk ranking and confirmation list.
Shared screening & optimization	Decision support & experiment design	`DS-DEC-001` Multi-model anti-infective triage	Mechanistic / rule-based analysis	Compound list, antiviral/antibacterial branch, activity/window gates and target species or protein.	Prioritize / review / defer lists, endpoint estimates, domain flags, reasons and CSV.
Shared screening & optimization	Decision support & experiment design	`DS-DEC-002` Combination reference surfaces and synergy analysis	Combination-effect analysis	Single-agent IC50 and Hill slopes, concentration grid, measured combination inhibition; MIC checkerboards require MIC data.	8×8 expected inhibition surfaces, observed deviations and equivalent-dose indices, interpreted by endpoint.
Shared screening & optimization	Decision support & experiment design	`DS-DEC-003` Next-batch experiment selection	Mechanistic / rule-based analysis	Embedded 500-compound antiviral or 1,000-compound antibacterial pool, weights and batch size; client pools need features and predictions.	Compound batch, selection rationale, activity–uncertainty plot and CSV.
Shared screening & optimization	Decision support & experiment design	`DS-DEC-004` Antiviral exposure and resistance simulation	Exploratory simulation	EC50, relative exposure, half-life, interval, mutation parameters and resistance fold change.	Viral load and resistant-fraction trajectories, 72-scenario scan and resistance-dominance timing.
Shared screening & optimization	Host-mechanism support	`DS-MOA-001` Host-protein modification motif scan	Mechanistic / rule-based analysis	Protein FASTA or amino-acid sequence.	Candidate sites, sequence context, motif scores and site map.
Shared screening & optimization	Custom modeling & validation	`DS-CUS-001` Biochemical-to-cellular efficacy calibration	Custom development · feasibility assessment	Structures, paired IC50/EC50, CC50, cell line and assay context; optional permeability/protein-binding data.	Calibrated model, error/domain report and cellular estimates only after prespecified gates are met.
Shared screening & optimization	Custom modeling & validation	`DS-CUS-002` New-target and new-species model development	Custom development · feasibility assessment	Target, endpoint, authorized structure–activity data and an independent test set.	Feasibility decision, model card, training/validation versions, prediction-interface specification or stop recommendation.
Shared screening & optimization	Custom modeling & validation	`DS-CUS-003` Independent validation and assay feedback	Custom development · feasibility assessment	Frozen model, training provenance and membership, external or prospective assay data, budget and prespecified metrics.	Independent performance report, enrichment/recall, error analysis, out-of-domain behavior and go/revise recommendation.

03 / INTEGRATED SERVICES

Integrated services for your research goals

Combine computational analysis and assay planning around your candidate library, available data and validation goals. Scope, timelines and deliverables are agreed at project initiation.

SERVICE / 01

Antiviral lead triage

For NA/PA-focused libraries progressing toward biochemical and cellular confirmation.

Receive target-specific activity rankings, cytotoxicity-proxy analysis, applicability tiers and a focused confirmation list.

AV-TGT-001 · AV-TGT-002 · DS-SAF-001 · DS-DEC-001

SERVICE / 02

Four-species antibacterial panel

For species-specific hit discovery and comparative antibacterial profiling.

Receive a compound-by-species MIC matrix, cytotoxicity proxies and a confirmation list, with separate species assessments for broad-spectrum objectives.

AB-MIC-001 · AB-MIC-002 · AB-MIC-003 · AB-MIC-004 · DS-SAF-001

SERVICE / 03

Lead optimization & next-batch design

For hit series balancing activity, developability and structural diversity.

Receive a five-endpoint developability matrix, a prioritized next batch, selection rationale and recommended follow-up assays.

DS-SAF-002 · DS-SAF-003 · DS-SAF-004 · DS-SAF-005 · DS-SAF-006 · DS-DEC-003

SERVICE / 04

Cellular-efficacy calibration pilot

For programs with paired measurements seeking more efficient cellular testing.

Establish data feasibility, a calibration baseline and independent holdout performance to define applicability and guide subsequent compound testing.

AV-SCR-002 · AV-DYN-001 · DS-CUS-001 · DS-CUS-003

04 / STUDY DELIVERY

From research goals to testable candidates

01 · Define the study

Define the target, species or cell line, measurement endpoints and testing capacity.

Deliverable: project scope and data submission template.

02 · Assess data quality

Review structures, duplicates, labels, batches and model coverage.

Deliverable: data-quality and applicability assessment with recommendations for missing measurements.

03 · Prioritize candidates

Integrate endpoint estimates and supporting evidence to distinguish priorities from candidates requiring additional measurements.

Deliverable: candidate rankings, evidence cards and applicability statements.

04 · Validate independently

Evaluate ranking utility on independent measured data and compare with baseline methods.

Deliverable: hit enrichment, recall, errors and assessment against agreed project criteria.

05 · Refine with assay feedback

Use positive, negative and failed results to guide the next screening round or model update.

Deliverable: confirmation recommendations and a model-update or expansion plan.

05 / PERFORMANCE & APPLICABILITY

Transparent performance. Defined applicability.

Interpret candidate rankings alongside development benchmarks, applicability assessment and independent project validation, accounting for assay, strain and chemical-space differences.

Match the intended endpoint

Biochemical potency, cellular efficacy and infection readouts are assessed separately. Cross-endpoint estimates require calibration with matched experimental data.

Align units and assay context

MIC results use consistent molar and mass units. Selectivity analysis aligns cell systems and assay conditions; cross-source ratios support preliminary triage.

Evaluate performance at each level

Computational consistency, model generalization and experimental efficacy are evaluated separately to define the decisions each result can support.

Measure value through confirmed hits

Evaluate hit enrichment, recall, cost per confirmed hit and applicability coverage at the intended testing capacity.

Development benchmarks for 12 activity and developability endpoints

n is source n_train; endpoint samples may overlap and cannot be summed into unique compounds. ρ is Spearman rank correlation; RMSE is in each endpoint’s target units.

CAT / Endpoint	n	ρ	RMSE	R²	Target units
`AV-TGT-001` NA inhibitor prioritization	1,143	0.809	1.070	—	−log10(mol/L)
`AV-TGT-002` PA endonuclease inhibitor prioritization	252	0.734	0.728	—	−log10(mol/L)
`AB-MIC-001` E. coli MIC prioritization	31,676	0.719	0.720	0.5537	−log10(mol/L)
`AB-MIC-002` K. pneumoniae MIC prioritization	12,015	0.754	0.701	0.5939	−log10(mol/L)
`AB-MIC-003` P. aeruginosa MIC prioritization	20,022	0.698	0.676	0.4888	−log10(mol/L)
`AB-MIC-004` S. aureus MIC prioritization	48,610	0.714	0.731	0.5283	−log10(mol/L)
`DS-SAF-001` Pooled cytotoxicity and selectivity reference	33,641	0.704	0.476	—	−log10(mol/L)
`DS-SAF-002` Aqueous-solubility triage	12,930	0.599	1.898	0.3454	log10(mol/L)
`DS-SAF-003` hERG inhibition triage	9,716	0.606	0.704	0.3889	−log10(mol/L)
`DS-SAF-004` CYP3A4 inhibition triage	5,490	0.573	0.652	0.411	−log10(mol/L)
`DS-SAF-005` CYP2D6 inhibition triage	2,993	0.579	0.695	0.3246	−log10(mol/L)
`DS-SAF-006` Human liver microsomal stability triage	601	0.449	0.488	0.1869	log10(min)

Benchmarks are reported from existing model records and have not been independently revalidated for this publication. MIC/ADMET figures refer to scaffold CV of 300-tree parent models; deployed subsets require separate confirmation. NA/CC50 use q-RASAR and PA uses Ridge + q-RASAR. These are not prospective efficacy results; RMSE is not a candidate-level confidence interval.

06 / YOUR RESEARCH PROJECT

Plan your research project

Select services and add your research goals to prepare a project brief for discussion. Details remain in this browser; export and share the brief when ready. Nothing is submitted automatically.

Research goal / target / species / cell lineAvailable input data

For research use only · Service scope and deliverables are defined for each project.

Online Inquiry

Submit your project details below, and our team will respond within 24 hours.

Free

Project Feasibility Assessment

Free

Project Sample

Need help getting the data you need?

Talk to our technical team about your project!

I Want To Talk