Flagship application

SFT Molecular Platform.

An integrated computational platform for molecular property prediction, candidate prioritisation, scientific validation context and reproducible research workflows.

The platform brings shared molecular representations, endpoint-specific prediction modules, validation context and scientific reporting into one connected environment. It is designed to help researchers move from molecular input to reviewable computational evidence while keeping experimental validation central to the discovery process.

Explore the SFT Molecular Platform Explore the workflow View platform releases
SFT
pKa logP Solubility BBB ADMET Binding
One connected environment

More than six separate prediction tools.

The SFT Molecular Platform is Spacefield Transformation Technologies Ltd's flagship computational application for molecular science and drug discovery research. It integrates shared molecular representations, endpoint-specific prediction models, controlled validation context and reproducible scientific reporting within one connected environment.

Its purpose is to reduce fragmentation across molecular prediction workflows. Researchers can move from common molecular input handling and shared computational representations to endpoint-specific models, candidate comparison and consistent scientific outputs without relying on disconnected tools.

The platform is being developed for academic and computational researchers, CRO and research teams, biotechnology companies and pharmaceutical R&D organisations that need prediction workflows that remain traceable, benchmarkable and suitable for integration into wider scientific programmes.

Academic & computational researchersSingle-molecule analysis, hypothesis development, candidate comparison and reproducible computational studies.
CRO and research teamsBatch prediction, shared reporting, candidate prioritisation and preparation for experimental evaluation.
Biotechnology & pharmaceutical R&DIntegrated prediction services, validation context and governed computational workflows for molecular research.
Platform workflow

From molecular input to reviewable scientific evidence.

A connected workflow standardises molecular input, generates shared computational representations, applies endpoint-specific prediction models and returns structured outputs with validation context for scientific review.

01

Molecular input

Submit a SMILES string, molecular record or supported batch dataset through a defined interface.

↓
02

Standardisation & descriptors

Molecular inputs are standardised and reproducible computational descriptors are generated.

↓
03

SFT-Core

Shared structure-aware and operator-inspired representations provide a common computational foundation.

↓
04

Prediction modules

Endpoint-specific models estimate molecular properties and generate comparative computational evidence.

↓
05

Validation context

Model identity, benchmark context, confidence information, scope and limitations remain attached to results.

↓
06

Scientific outputs

Predictions, rankings, reports and reproducibility artefacts support interpretation and experimental planning.

Integrated prediction modules

Six endpoint technologies within one platform.

Each module addresses a distinct molecular question while using shared platform services for input handling, descriptors, model execution, reporting and audit.

pKaDevelopment

SFT-pKa

Predicts molecular ionisation behaviour and pKa-related properties using structure-aware computational representations and validated statistical modelling. Supports physicochemical analysis, molecular profiling and downstream prediction workflows.

Supports prediction of molecular pKa-related behaviour using structure-aware descriptors and validated statistical modelling.

Primary output
Predicted pKa values and contextual model information
Research use
Ionisation analysis, molecular comparison and downstream property interpretation
logPDevelopment

SFT-logP

Predicts molecular lipophilicity and partitioning behaviour from molecular structure. Supports assessment of physicochemical balance, permeability-related behaviour and comparative candidate profiling.

Estimates partition behaviour through molecular descriptors designed to capture hydrophobic, polar and structural contributions.

Primary output
Predicted logP and associated reporting artefacts
Research use
Compound comparison, permeability context and formulation-oriented analysis
H₂ODevelopment

SFT-Solubility

Predicts aqueous solubility from molecular structure and related physicochemical features. Supports early evaluation of molecular developability, formulation risk and comparative compound assessment.

Combines molecular structure, polarity, ionisation context and learned relationships to estimate aqueous behaviour.

Primary output
Predicted solubility with structured result records
Research use
Early developability assessment and candidate prioritisation
BBBDevelopment

SFT-BBB

Predicts blood-brain barrier permeability using molecular structure, polarity, lipophilicity and related transport-informed features. Supports prioritisation of compounds for central nervous system research.

Evaluates structural and physicochemical patterns associated with central nervous system permeability.

Primary output
BBB class or probability-based prediction
Research use
CNS-oriented screening and comparative permeability assessment
ADMEResearch

SFT-ADMET

Supports computational assessment of absorption, distribution, metabolism, excretion and toxicity-related properties using endpoint-specific models. Provides early developability context for candidate comparison and research prioritisation.

Provides a multi-endpoint framework for evaluating selected pharmacokinetic and safety-related molecular properties.

Primary output
Endpoint predictions, summaries and comparative records
Research use
Multi-parameter review and early risk-oriented prioritisation
KdValidation

SFT-Binding

Supports target-specific protein-ligand affinity modelling using ligand, protein and interaction-aware computational representations. Helps compare and prioritise candidate compounds for further experimental evaluation.

Combines ligand descriptors, protein representations and interaction-aware modelling for binding prediction workflows.

Primary output
Predicted affinity or ranked ligand–target evidence
Research use
Target-specific candidate comparison and computational screening

Scientific position: Computational predictions generated by the SFT Molecular Platform support molecular analysis, candidate prioritisation and hypothesis generation. They do not replace chemical synthesis, laboratory testing, biological assays, safety studies or independent experimental validation.

Unified platform architecture

Shared foundations, endpoint-specific intelligence.

The architecture keeps common platform functions central while allowing each molecular endpoint to retain specialised descriptors, models and validation records.

Standardised molecular input ↓
SFT-CoreShared molecular representations and platform services
│
SFT-pKaIonisation
SFT-logPLipophilicity
SFT-SolubilityAqueous behaviour
SFT-BBBPermeability
SFT-ADMETMulti-endpoint properties
SFT-BindingLigand–target affinity
│
Unified reporting engineConsistent CSV, JSON, PDF, logs, manifests and integration outputs
↓ Traceable scientific reports and workflow artefacts
Scientific outputs

Structured outputs for scientific review and integration.

The platform is designed to return more than individual predictions. It produces structured outputs for interpretation, benchmarking, reproducibility and integration into wider research workflows.

01

Prediction outputs

  • Endpoint predictions
  • Candidate rankings
  • Confidence or applicability context
02

Scientific reports

  • PDF scientific reports
  • CSV result datasets
  • JSON records
03

Validation records

  • Benchmark summaries
  • Model-performance context
  • Validation, runtime and audit logs
04

Reproducibility artefacts

  • Frozen manifests
  • Integrity hashes
  • Dataset fingerprints
  • Configuration snapshots
05

Integration outputs

  • REST API responses
  • SDK exports
  • Batch and CLI reports
  • Workflow artefacts
Example research workflow

A clear path from SMILES to reviewable evidence.

A typical workflow validates the molecular input, generates descriptors, executes selected prediction modules and produces a structured scientific record.

The user can run a single endpoint or a multi-module workflow depending on the scientific question. Results remain linked to the model, configuration and output artefacts used during execution.

SMILESMolecular input
↓
ValidationInput standardisation and checks
↓
Descriptor engineShared and endpoint-aware features
↓
Prediction modulesSelected molecular endpoints
↓
Confidence analysisModel context and comparative evidence
↓
Scientific reportTraceable outputs and workflow artefacts
Why the platform is different

Computational prioritisation before experimental commitment.

The platform is designed to support experimental decision-making, not replace it. Its value lies in helping teams organise evidence, compare candidates and focus laboratory effort.

Broad experimental workflow
Large candidate space
↓
Experimental screening
↓
High resource demand
↓
Longer iteration cycles
VS
SFT-supported workflow
Mechanism-informed descriptors
↓
Computational prioritisation
↓
Candidate ranking and evidence
↓
Focused experimental validation

This comparison describes the intended workflow role of the platform. It does not establish a universal reduction in cost, time or experimental burden; those outcomes depend on the specific project, data and validation design.

Platform roadmap

A controlled path from research software to production release.

Platform development follows a defined maturity lifecycle: Research → Development → Validation → Release Candidate → Production Release. Individual modules may progress through this lifecycle at different rates as benchmarking, integration and release verification are completed.

Current development state

Integrated platform foundation

  • SFT-pKa — Development
  • SFT-logP — Development
  • SFT-Solubility — Development
  • SFT-BBB — Development
  • SFT-ADMET — Research
  • SFT-Binding — Validation
  • Shared reporting and reproducibility artefacts
Development direction

Unified research and deployment environment

  • Unified molecule API
  • Web-based research interface
  • Python SDK and developer documentation
  • Large-scale batch processing
  • REST API services
  • Interactive scientific dashboards
  • Organisation-ready deployment and support workflows
Continue exploring

Technology, evidence and releases.

Explore how Spacefield builds its computational architecture, how platform performance is examined and how software releases are prepared for reproducible use.

Explore the SFT Molecular Platform Explore technology View research Platform releases Contact Spacefield