Formulation, deformulation, optimisation, freedom to operate, patent drafting, VCM process design, literature review and data analysis.
Each study answers from DeepC’s library and your own documents, and returns what formulation scientists work with: QTPP01 profiles, F01 candidates, DF01 reconstructions, F02A variants, FTO01 opinions, PA01 patent templates, VCM01 process designs and analysis reports.
The research studies take on the questions a development team meets between preformulation and manufacture.
One study’s deliverable starts the next where the science does: a freedom-to-operate opinion becomes a patent template, a reconstructed reference product becomes targeted variants, a formulation becomes an Optimize campaign.
After a freedom-to-operate analysis, one button starts patent application drafting from the lead candidate.
A formulation optimization can start from a reconstructed reference product.
Open in Optimize restates a candidate as the target of an experimental campaign.
Each with the question it is asked, the stages it works through, and what it hands back.
Designs the formulation, starting from a QTPP.
ICH Q8(R2), CTD 3.2.P.2
Procedure
Records FRM-0001 onwards, numbered in your organisation’s own sequence.
Every design opens with a Quality Target Product Profile, set out before any candidate is proposed. The report follows ICH Q8(R2) and CTD Module 3.2.P.2, with a risk assessment and a control strategy, and each candidate weighs the API, its excipients, the patent landscape and how it will manufacture at scale.
What it returns
A QTPP, then 3 to 5 candidate formulations (F01, F02…), each from a different angle and ranked against the others.
“Design an oral formulation for this poorly soluble compound.”
Improves a formulation you already have.
Procedure
Records OPT-0001 onwards, numbered in your organisation’s own sequence.
Starts from an existing formulation, one DeepC designed or a reconstructed reference product, and returns variants anchored to it. It can pursue several goals at once, such as lower cost with better dissolution and a simpler process, and states plainly what each variant gives up for what it gains. For a campaign that learns from your bench results round by round, DeepC Optimize takes over.
What it returns
Up to three variants of a parent (F02 → F02A, F02B, F02C), each with what changed, a before-and-after comparison, its trade-offs and the next experiments to run.
“Optimise this tablet for 20% cost reduction while maintaining dissolution.”
Synthesises the evidence across every source.
Procedure
Records LIT-0001 onwards, numbered in your organisation’s own sequence.
For questions that need synthesis across literature, patents, clinical trials, regulatory documents and toxicology, a literature review runs a broader investigation and writes it up in prose: comparing formulation approaches, surveying a technology landscape, assessing feasibility and reading regulatory precedent. Candidate formulations come from a formulation study.
What it returns
A review-style synthesis in prose, every claim cited.
“Compare lipid-based and amorphous solid dispersion approaches for BCS Class II compounds.”
Reconstructs a reference product as a Q1/Q2 estimate.
Procedure
Records DEF-0001 onwards, numbered in your organisation’s own sequence.
Starts from a marketed or reference product and reconstructs its most plausible quantitative composition from public evidence, grading its confidence in every estimate and naming every gap. It can follow with up to three candidate formulations of your own.
What it returns
A deformulation report (DF01): the best-estimate Q1/Q2 composition, the dissolution target, the patents around the product, where each ingredient can be sourced, and a graded confidence on every estimate.
“Deformulate the reference product for Drug X 500 mg tablets and estimate its Q1/Q2 composition.”
Writes a freedom-to-operate opinion before you commit.
Procedure
Records FTO-0001 onwards, numbered in your organisation’s own sequence.
Maps the patent landscape before you commit to an approach, rates the risk and proposes ways around the patents in your path. For a generic it sets out a filing strategy against the reference product’s listed patents, and it weighs the risk market by market. A Draft Patent Application button then starts patent application drafting from the lead candidate.
What it returns
An FTO opinion (FTO01): claim charts with the claims in their own words, prior art, an overall risk rating, design-arounds, an expiry timeline and recommendations; then three design-around formulations.
“Analyse FTO for an HPMC-based extended-release tablet for Drug X.”
Drafts patent application templates for your attorney.
Procedure
Records PAT-0001 onwards, numbered in your organisation’s own sequence.
Drafts stream into the side panel as DeepC writes, so you can watch the document take shape and redirect early. The template carries the full specification, from the field and related art to the examples, the numbered claims and the abstract. It never fabricates experimental data: stability, dissolution and comparative results are marked as placeholders for your laboratory.
What it returns
A US nonprovisional application template (PA01) for your patent attorney to complete and refine before filing.
“Draft a patent application template for F02 with three independent claims.”
Designs amorphous solid dispersion screens on VCM.
Procedure
Records VCM-0001 onwards, numbered in your organisation’s own sequence.
Built for MeltPrep Vacuum Compression Molding, including VCM as a material-sparing pre-screen before hot-melt extrusion scale-up. It sets processing parameters from drug-polymer compatibility, checks the risk of thermal degradation, designs screens that use a few grams of material in total, and troubleshoots the defects a disc can show.
What it returns
VCM process designs (VCM01, VCM02…): drug and polymer, process parameters, the protocol, the expected outcome, the characterisation to run, and warnings.
“Design a VCM screen for itraconazole on Kollidon VA 64 and HPMCAS-L.”
Analyses your data and writes the report.
For work beyond a quick calculation, such as model training, long runs or a formal report, DeepC takes your data into an isolated workspace of its own, plans the analysis, runs it and writes the report with publication-quality figures. Results stay with the conversation after the run ends.
What it returns
A report with figures, exported as PDF or Word; the figures, tables and code arrive as files in the conversation.
“Run a DoE-matrix analysis on this dissolution dataset and report optimal parameter ranges.”
Biologics suite
Biologics research review, FTO analysis, and formulation and CMC study, for antibodies, ADCs, bispecifics, mRNA-LNP, AAV and cell therapy.
Safety suite
ERA, extractables, leachables, nitrosamines, OEL, and PDE, anchored to ICH M7, EMA HBEL, USP <1664>, and ECETOC TR 101.
Overview
What DeepC hands back, how it shows its work, and the workspace around it: the platform overview.
From a first call to a pilot on your own programme.
A call with a formulation lead, who is in touch within a business day.
On your own programme, scoped with your team.
The AI model providers behind DeepC retain none of your content and train nothing on it.
Prefer email? info@deepceutix.com