MendelFOLD

University spin-off developing computational methods for biorelevant protein 3D structure and function prediction.

Cover Block

Open sources

Field	Value
Name	MendelFOLD
Tagline	University spin-off developing computational methods for biorelevant protein 3D structure and function prediction [Lupa.cz, December 2025]
Headquarters	Brno, Czech Republic [LinkedIn]
Founded	2024 [LinkedIn]
Stage	Pre-Seed
Business Model	B2B
Industry	Deeptech
Technology	Biotech / Life Sciences
Geography	Eastern Europe
Growth Profile	Venture Scale
Founding Team	Academic Spinout
Funding	Undisclosed [InnovPrecMed, October 2025]

What an Investor Needs First

PUBLIC MendelFOLD is a 2024 university spin-off in Brno developing computational methods for protein 3D structure and function prediction, and it merits investor attention because it is trying to commercialize a distinct scientific thesis in a market where better prediction tools could matter for therapeutics and synthetic biology [Lupa.cz, December 2025] [EIT Higher Education Initiative / Falling Walls Ventures, May 2026]. The company appears to have emerged from research at Mendel University’s Institute of Chemistry and Biochemistry, with incorporation reported on November 27, 2024, positioning it as a recent academic commercialization rather than a mature software company [Lupa.cz, December 2025] [PERPLEXITY SONAR PRO BRIEF].

Its product claim is not a general-purpose AI platform so much as a knowledge-based protein prediction approach rooted in what its materials describe as a proteomic code, with public program materials also pointing to atomic-level structure-function prediction methods for synthetic biology and therapeutics [EIT Higher Education Initiative / Falling Walls Ventures, May 2026] [The proteomic code - Brno, May 2025] [The proteomic code: Novel amino acid residue pairing models "encode" protein folding and protein-protein interactions - PubMed, 2026]. The relevant differentiation, if it holds up experimentally, rests on the scientific framework and underlying residue-pairing models rather than on scale claims, customer traction, or disclosed commercial benchmarks, none of which are yet established in public reporting [InnovPrecMed, October 2025] [PERPLEXITY SONAR PRO BRIEF].

The founding bench is unusually academic: Andrew D. Miller is publicly identified as a Professor of Organic Chemistry and Chemical Biology at Mendel University in Brno, Zbyněk Heger leads the university’s Department of Chemistry and Biochemistry, and Tareq Yakoub Hassan Hameduh is described as a postdoctoral researcher in computational protein science [LinkedIn] [doc. Mgr. Zbyněk Heger, Ph.D. - Brno - Ústav techniky, 2026] [PERPLEXITY SONAR PRO BRIEF]. That profile gives the company scientific depth early, but it also means investors will likely spend the next 12 to 18 months looking for evidence that the science can convert into repeatable product performance, external validation, and commercial relationships.

On financing, the public record is still thin: IP Lab Ventures is associated with the company through venture-building and commercialization activity, and a first significant funding milestone was reported in September 2025 without public disclosure of size, instrument, or lead investor [European Investment Fund, 2026] [IP Lab Ventures] [PERPLEXITY SONAR PRO BRIEF]. The business model is B2B and the most plausible initial route appears to be collaboration-driven work with pharmaceutical and synthetic-biology companies, but the near-term watch items are straightforward: independent proof that the prediction framework works better on biorelevant problems, named partnerships or pilots, and cleaner visibility into capitalization and go-to-market execution [PERPLEXITY SONAR PRO BRIEF] [InnovPrecMed, October 2025].

Reasoned from indirect evidence -- This section relies on a mix of independent program and media sources, plus several material claims that are only singly sourced or surfaced through compiled research.

Taxonomy Snapshot

Axis	Value
Stage	Pre-Seed
Business Model	B2B
Industry / Vertical	Deeptech
Technology Type	Biotech / Life Sciences
Geography	Eastern Europe
Growth Profile	Venture Scale
Founding Team	Academic Spinout
Funding	Undisclosed

Inside the Company

PUBLIC MendelFOLD is a Brno-based university spin-off working on computational protein structure and function prediction, with its public footprint still closer to a research commercialization effort than a scaled operating company [LinkedIn]. The company was incorporated as MendelFOLD s.r.o. on November 27, 2024, in the Czech Republic, and public registry details surfaced in the source set indicate registered capital of CZK 20,000 and directors Andrew Hladký and Nicholas Geoffrey Alan Weaver [LinkedIn].

The founding story matters because nearly all of the company’s public positioning flows from the university lab rather than from commercial traction. Public materials tie the company to research at Mendel University’s Institute of Chemistry and Biochemistry, and investor ecosystem materials from IP Lab Ventures and the European Investment Fund place MendelFOLD among deeptech spin-offs emerging from that commercialization network [IP Lab Ventures] [European Investment Fund, 2026].

Chronology is sparse but coherent in the available filings and institutional announcements. The company appears to have formed in late 2024, then surfaced more visibly through technology-transfer and venture-program channels in 2025, including InnovPrecMed’s October 2025 notice that it had been nominated for Falling Walls Venture [InnovPrecMed, October 2025]. By May 2026, EIT Higher Education Initiative and Falling Walls Ventures were publicly profiling MendelFOLD’s work in protein folding and structure-function prediction, which suggests the company had moved from incorporation into early ecosystem validation, even if product, customer, and financing disclosures remain limited [EIT Higher Education Initiative / Falling Walls Ventures, May 2026].

Partially corroborated -- Built from LinkedIn and institutional ecosystem sources, with legal entity details cited from the public source set but not independently corroborated here by a state filing link.

Under the Hood

Core approach

MIXED MendelFOLD is making an unusually specific scientific claim for a company this early: that protein structure and function prediction can be improved by a knowledge-based framework rooted in what it describes as a "proteomic code," rather than relying only on pattern recognition from large training datasets [Lupa.cz, December 2025] [InnovPrecMed, October 2025]. Public descriptions are consistent on the practical problem being addressed, namely computational prediction of biorelevant protein 3D structure and function for applications in drug discovery, synthetic biology, and therapeutics, but they are much thinner on product surface, workflow, and user experience [EIT Higher Education Initiative / Falling Walls Ventures, May 2026] [European Investment Fund, 2026]. That matters because the external evidence supports a research platform thesis more clearly than a finished software-product thesis.

The technical narrative visible in public sources ties the company to atomic-level molecular simulation and to a body of academic work around amino acid residue pairing, protein folding, and the proposed proteomic code framework [EIT Higher Education Initiative / Falling Walls Ventures, May 2026] [The proteomic code: Novel amino acid residue pairing models "encode" protein folding and protein-protein interactions - PubMed, 2026] [The proteomic code - Brno, May 2025]. The company's differentiation, to the extent it can be assessed from public evidence, appears to rest on whether that scientific framework produces better biologically relevant predictions than purely data-driven alternatives in real R&D settings. Public materials do not yet disclose benchmark results, validated case studies, or named commercial deployments, so the scientific proposition is legible while the degree of productization remains unproven [InnovPrecMed, October 2025] [EIT Higher Education Initiative / Falling Walls Ventures, May 2026].

Evidence base and current maturity

MIXED The strongest public signal is not a product demo or customer implementation, but institutional validation around origin and intended use. MendelFOLD is described as a university spin-off from Mendel University in Brno, and program materials linked to InnovPrecMed and Falling Walls Ventures position it as translating academic research into computational tools for synthetic biology and next-generation therapeutics [Lupa.cz, December 2025] [InnovPrecMed, October 2025] [EIT Higher Education Initiative / Falling Walls Ventures, May 2026]. That is constructive, but it is still one step removed from proof that the platform has crossed into repeatable commercial use.

The academic record does at least show that the underlying theory is being articulated in recent publications associated with Andrew D. Miller and collaborators, including work on residue pairing models, folding information in nucleic acids, and phylogenomic analyses of dipeptides across large proteome datasets [The Proteomic Code: A molecular recognition code for proteins, 2026] [The proteomic origin of the genetic code: Expert Review of Proteomics: Vol 23, No 3, 2026] [Full article: The proteomic origin of the genetic code, 2026]. Even so, investors should separate publication activity from product validation. Based on public evidence alone, MendelFOLD today reads as a science-first computational platform with a clearly defined hypothesis and plausible pharma and synthetic-biology relevance, but with limited disclosed evidence on performance, delivery model, or integration into customer workflows [InnovPrecMed, October 2025] [EIT Higher Education Initiative / Falling Walls Ventures, May 2026].

Partially corroborated -- Core company description is corroborated across Lupa.cz, InnovPrecMed, Falling Walls Ventures, and the European Investment Fund, but several deeper technical claims rely on company-adjacent or founder-linked materials rather than independent benchmarking.

Market Research

PUBLIC

The market matters now because protein-structure prediction has moved from a narrow computational biology problem to a practical input for drug discovery, synthetic biology, and platform biotech, but the commercial opening remains uneven and still favors tools that can show biological relevance rather than headline model performance alone [EIT Higher Education Initiative / Falling Walls Ventures, May 2026].

Public source coverage for MendelFOLD does not include a disclosed TAM, SAM, or SOM, and the available materials do not point to a named third-party market study specific to the company’s target wedge [InnovPrecMed, October 2025]. That limits precision. Still, the company’s stated application areas, synthetic biology and therapeutics, place it in the overlap between computational biology software, structure-based drug discovery, and research tools sold into pharma and academic-industry translational programs [EIT Higher Education Initiative / Falling Walls Ventures, May 2026]. On the public record, the cleaner read is that MendelFOLD is pursuing a narrow enabling-tools position inside a much larger life sciences R&D budget pool, not a standalone end-market that can yet be sized from disclosed company data [Lupa.cz, December 2025].

The demand side is easier to see than the market size. MendelFOLD’s pitch rests on a claim that existing purely data-driven AI methods have limitations for biologically relevant structure and function prediction, and that a knowledge-based approach grounded in molecular principles may improve usefulness in research settings [PERPLEXITY SONAR PRO BRIEF]. External program descriptions broadly align with that framing: InnovPrecMed positions the company within precision-medicine commercialization and partnership formation, while the Falling Walls profile places the work closer to atomic-level simulation and next-generation therapeutics than to generic AI software [InnovPrecMed, October 2025] [EIT Higher Education Initiative / Falling Walls Ventures, May 2026]. That suggests demand, if it materializes, will likely be buyer-specific and validation-heavy, with pharma research teams and synthetic-biology groups testing whether the output is actionable in experimental workflows.

Adjacent markets matter here because buyers may compare MendelFOLD less against a single direct peer than against substitute approaches already embedded in discovery stacks. Those substitutes include conventional molecular modeling, broader bioinformatics platforms, AI-native protein design systems, and outsourced discovery services, all of which compete for the same R&D attention and budget even when the technical methods differ [EIT Higher Education Initiative / Falling Walls Ventures, May 2026]. The company’s own scientific framing, centered on a proposed proteomic code and amino-acid residue pairing models, also places it near translational academic research markets where credibility often depends on publication uptake, reproducibility, and partner-generated evidence before software revenue scales meaningfully [The proteomic code - Brno, May 2025] [The proteomic code: Novel amino acid residue pairing models "encode" protein folding and protein-protein interactions - PubMed, 2026].

Macro and regulatory forces are supportive in a general sense, but the public evidence supports only a measured claim. European research-commercialization programs are clearly part of the company’s environment: InnovPrecMed and EIT-linked activity indicate policy support for technology transfer, entrepreneurship, and academic-industry collaboration in precision medicine and related fields [InnovPrecMed, October 2025] [EIT Higher Education Initiative / Falling Walls Ventures, May 2026]. That can help early validation and network access. It does not, on the available record, establish reimbursement pathways, regulatory approvals, or procurement demand. For a company positioned as an upstream discovery tool rather than a regulated therapeutic product, the nearer-term market constraint is likely scientific adoption and partner proof rather than formal product regulation [Lupa.cz, December 2025].

Market lens	Publicly supported observation	Source
Core target domain	Protein structure and function prediction for synthetic biology and therapeutics	[EIT Higher Education Initiative / Falling Walls Ventures, May 2026]
Commercial route implied	Collaborations with pharmaceutical and synthetic-biology companies	[PERPLEXITY SONAR PRO BRIEF]
Institutional tailwind	Supported by programs focused on precision medicine, technology transfer, and partnerships	[InnovPrecMed, October 2025]
Market posture	University spin-off commercializing research from Mendel University	[Lupa.cz, December 2025]

The table shows why this is best understood as a market-access story, not yet a market-size story. Public evidence supports the problem area, buyer classes, and institutional tailwinds, but not a quantified revenue pool or adoption curve specific to MendelFOLD.

Partially corroborated -- Section relies on program materials and media coverage that corroborate the company’s target domains and commercialization context, but no independent third-party market sizing report or disclosed company metrics were available.

Competition and Substitutes

MIXED MendelFOLD appears to be positioning itself against mainstream protein-structure prediction and molecular-design workflows by arguing that a knowledge-based, biology-first approach can address limits in purely data-driven methods, but the public record does not yet identify direct named competitors in the company’s own materials [Lupa.cz, December 2025] [EIT Higher Education Initiative / Falling Walls Ventures, May 2026].

The competitive map is still readable even without a disclosed peer set. On one side sit incumbent discovery workflows inside large pharmaceutical companies and research institutes, where protein modeling is often embedded in broader wet-lab and computational pipelines rather than sold as a standalone product. On another sit AI-first structure and design platforms that frame prediction as a scale problem for models and training data. A third bucket is adjacent substitute infrastructure: academic software, internal bioinformatics teams, and contract research collaborations that can absorb some of the same budget line even if they do not present themselves as direct vendors. MendelFOLD’s public materials suggest it wants to sell into pharmaceutical and synthetic-biology use cases, which places it in competition not just with specialist prediction tools but with any workflow that already produces usable target hypotheses for those buyers [InnovPrecMed, October 2025] [EIT Higher Education Initiative / Falling Walls Ventures, May 2026].

Its clearest edge today is scientific distinctiveness, not distribution. The company traces its platform to a claimed "proteomic code" and related residue-pairing models associated with Andrew D. Miller and collaborators, with recent papers arguing that folding information is encoded in nucleic-acid and amino-acid relationships rather than inferred only from large data corpora [The proteomic code - Brno, May 2025] [The proteomic code: Novel amino acid residue pairing models "encode" protein folding and protein-protein interactions - PubMed, 2026] [The proteomic origin of the genetic code: Expert Review of Proteomics: Vol 23, No 3, 2026]. That gives MendelFOLD a differentiated scientific story and a founder-market fit rooted in Mendel University research, reinforced by the academic profiles of Miller, Zbyněk Heger, and Tareq Yakoub Hassan Hameduh [Prof Andrew D. Miller FRSC CChem - Professor of Organic Chemistry and Chemical Biology at Mendel University in Brno | LinkedIn, 2026] [doc. Mgr. Zbyněk Heger, Ph.D. - Brno - Ústav techniky, 2026]. The durability of that edge is still uncertain. If the advantage depends on proprietary know-how, validated benchmarks, and patentable implementation, it could compound. If it depends mainly on a novel framing that larger model-driven groups can test and reproduce, it is perishable.

The company is most exposed where scale, validation, and buyer access matter more than originality. Public sources tie MendelFOLD to IP Lab Ventures and EIT-linked support, but they do not disclose a commercial customer base, benchmark outperformance, or a funded go-to-market motion that would let it win on speed of adoption [European Investment Fund, 2026] [IP Lab Ventures] [InnovPrecMed, October 2025]. That leaves the company vulnerable to better-capitalized platform players and to internal R&D teams at pharmaceutical companies, which can treat an unproven external method as one input among many rather than as a system of record. It also leaves MendelFOLD exposed to categories it cannot yet credibly enter from the public evidence alone, including end-to-end drug discovery platforms that combine target identification, wet-lab validation, and enterprise business development under one roof.

The most plausible 18-month scenario is a bifurcation between scientific credibility and commercial relevance. MendelFOLD is the most plausible winner if its approach can show reproducible gains on biologically relevant structure-function problems that standard model-centric pipelines miss, especially in synthetic biology or narrowly defined therapeutic programs where interpretability matters as much as raw prediction quality [EIT Higher Education Initiative / Falling Walls Ventures, May 2026]. MendelFOLD is also the most plausible loser if buyer behavior continues to consolidate around incumbent internal workflows and larger external platforms that can bundle prediction with downstream validation, service capacity, and enterprise procurement readiness. For now, the company’s competitive position rests more on a differentiated scientific hypothesis than on owned channels or demonstrated market power.

Partially corroborated -- Section relies on named public sources for MendelFOLD’s positioning, scientific basis, university origins, and support network, but no named competitors or independent commercial benchmarks were confirmed in the public record.

Opportunity

Upside case

PUBLIC The prize here is unusually large if the science translates: a working protein structure and function prediction platform that improves hit-finding or design decisions for drug discovery and synthetic biology could become a high-value upstream tool in two budgets that already support specialized computational infrastructure, even though MendelFOLD is still at an early, lightly disclosed stage [EIT Higher Education Initiative / Falling Walls Ventures, May 2026] [InnovPrecMed, October 2025].

The headline opportunity is not simply to be another protein-modeling startup. It is to become a differentiated discovery layer for cases where purely data-driven methods leave gaps, using a knowledge-based approach tied to what the company and affiliated researchers describe as a proteomic code and residue-pairing logic for structure and function prediction [Lupa.cz, December 2025] [The proteomic code: Novel amino acid residue pairing models "encode" protein folding and protein-protein interactions - PubMed, 2026]. That outcome is still conditional on technical proof, but it is at least reachable on the public record for three reasons: the company is a formal university spin-off from Mendel University research rather than a greenfield software project [Lupa.cz, December 2025] [European Investment Fund, 2026]; the scientific thesis appears to rest on a research program with antecedents going back to 2002 rather than a newly assembled marketing claim [The proteomic code - Brno, May 2025]; and external institutions linked to technology transfer and venture programming have already elevated the company within EIT-supported channels and the IP Lab orbit [InnovPrecMed, October 2025] [European Investment Fund, 2026]. In plain terms, if MendelFOLD can show that its methods improve decision quality in therapeutic or synthetic-biology workflows, the company could matter as infrastructure rather than as a one-off services shop.

The public evidence supports a few distinct paths to scale, each of which depends on a different proof point.

Scenario	What happens	Catalyst	Why it's plausible
Discovery engine for therapeutics	MendelFOLD becomes a specialist platform used by pharma and biotech teams to prioritize protein targets, folding hypotheses, or structure-function relationships in early discovery workflows	A first disclosed collaboration or validation program with a pharmaceutical partner	The company has publicly indicated an intention to pursue pharmaceutical collaborations, and external profiles place therapeutics among the intended applications [PERPLEXITY SONAR PRO BRIEF] [EIT Higher Education Initiative / Falling Walls Ventures, May 2026]
Design layer for synthetic biology	The platform is adopted by synthetic-biology groups that need protein design or functional prediction in faster iteration loops than wet-lab screening alone can provide	A partner use case showing reduced design cycles or improved candidate selection	Synthetic biology is named as an intended end market in public materials, and the company frames its work at the atomic structure-function level rather than as a generic AI workflow tool [PERPLEXITY SONAR PRO BRIEF] [EIT Higher Education Initiative / Falling Walls Ventures, May 2026]
Spin-off science to platform IP	MendelFOLD turns a research thesis into licensable computational IP that larger drug-discovery or platform players embed rather than rebuild internally	A patenting or IP-commercialization milestone tied to university transfer channels	InnovPrecMed explicitly describes support around IP strategy and partnerships, and IP Lab materials place the company inside a commercialization pipeline for university-derived deeptech [InnovPrecMed, October 2025] [IP Lab Ventures]

What compounding would look like is straightforward, even if it is not yet visible in operating metrics. In this model, each successful prediction program could generate more benchmark data on where the company's methods outperform incumbent approaches, which in turn strengthens the next BD conversation and narrows the product to the highest-value workflows. That is not a consumer-style network effect. It is a research-platform flywheel, where validation creates credibility, credibility attracts partners, and partner work creates harder-to-replicate know-how around which proteins, functions, and use cases the method handles best [EIT Higher Education Initiative / Falling Walls Ventures, May 2026] [InnovPrecMed, October 2025].

There is a second layer of compounding if the scientific premise proves durable. The research record around the proteomic code, residue-pairing models, and dipeptide phylogeny suggests the founders are trying to build from a distinct theoretical base rather than fine-tuning the same public modeling stack as everyone else [The proteomic code: Novel amino acid residue pairing models "encode" protein folding and protein-protein interactions - PubMed, 2026] [The proteomic origin of the genetic code: Expert Review of Proteomics: Vol 23, No 3, 2026] [Full article: The proteomic origin of the genetic code, 2026]. If that base yields reproducible performance in narrow but valuable domains, the moat may come less from scale data and more from proprietary interpretation layers, accumulated validation results, and the cost to replicate a long-running scientific program inside a buyer organization.

The size of the win is best framed as a scenario, not a forecast. Public materials here do not provide a defensible market-sizing figure or a clean disclosed peer set, so the conservative way to think about upside is strategic value rather than a modeled TAM. If MendelFOLD were to become a trusted computational layer for therapeutic discovery or synthetic-biology design, the company could plausibly support platform-level value as a scarce deeptech asset inside European biotech tooling, particularly because it sits at the intersection of university IP, computational biology, and commercialization support infrastructure [European Investment Fund, 2026] [InnovPrecMed, October 2025]. That is still several proof steps away from investability on fundamentals alone, but the public record is sufficient to say the upside case rests on owning a consequential point in the biology R&D stack, not on selling generic software.

Partially corroborated -- Section relies on institutional profiles, university-transfer coverage, and founder-linked scientific publications; core commercial outcomes remain unverified in public sources.

Sources