AI Drug Discovery
One nanobody targeting the MIF/CD74 axis across MS, ALS, and Brain Metastasis. Designed entirely in silico through AI foundation models (scGPT) and multi-scale Digital Twins.
The NeuroInteractome Platform
From a raw snRNA-seq dataset to a patent-ready .pdb file in weeks. Now includes trimer-aware structural validation via HDOCK against experimental structures. JFIN-0001 is the first output — the same engine targets Alzheimer's, Parkinson's, or any neuroinflammatory disease.
snRNA-seq
Transcriptomic ingestion
scGPT
Foundation model
Digital Twin
In silico perturbation
VHH Design
De novo nanobody
MD Simulation
OpenMM dynamics
ESM-2 Evolution
Directed evolution
Trimer Validation
HDOCK vs experimental structure
Convergent Mining
AIRR validation
Immunogenicity
MHC-II / Humanization
Developability
CMC profiling
INPUT
snRNA-seq dataset (.h5ad)
OUTPUT
Patent-ready nanobody (.pdb + full dossier)
Fully automated. Disease-agnostic. The same pipeline validated across RA (synovial), IBD (intestinal), COVID-19 (pulmonary), and MS (CNS) tissues with 100% retrospective drug rediscovery accuracy.
Lead Asset Family — JFIN Series
Starting from JFIN-0001 (116 aa, 12.6 kDa), our platform generated 48 trimer-aware next-generation candidates via ESM-2 directed evolution. Lead candidate validated by HDOCK rigid-body docking against the experimental CD74 trimer (PDB 1IIE). Patent filed OEPM April 2026; v2 addition in preparation.
The NeuroInteractome Digital Twin identified the MIF/CD74 axis as the dominant neuroinflammatory driver in MS microglia. JFIN-0001 blocks this axis at the extracellular domain level.
BBB Advantage vs IgG
VHH ~2% vs IgG ~0.1% brain penetration
H-bonds at Interface
Molecular dynamics (100 ps, 310 K)
Validated Target
scGPT + Digital Twin discovery
Structure Confidence
ESMFold prediction quality
Digital Twin simulation predicts remyelination via MIF/CD74 blockade in microglial-oligodendrocyte crosstalk.
Absolute Novelty confirmed: zero exact CDR3 matches across >1 billion sequences. Yet nature independently converges on the same binding chemistry.
Exact CDR3 Matches
Unique design — novel IP confirmed
Convergent Antibodies
Natural IGH sharing DFVYY motif
Top CDR3 Similarity
Human IGH encoded by IGHD3-22
Convergent T-Cell Receptors
Cross-arm validation (TRB chains)
The aromatic binding motif DFVYY is encoded by human germline D-gene IGHD3-22. Our AI rediscovered a solution selected by 500M years of evolution.
Grade A Humanization — 100% framework identity to human IGHV3-11 germline. Developability profile confirms manufacturability with standard VHH expression systems.
Framework Identity
IGHV3-11 (Grade A, near-human)
Humanness Score
Top 20 germline avg identity
R103F Isoelectric Point
Optimal range for high solubility
Instability Index
< 40 = stable. Excellent score.
R103F variant: reduced MHC-II epitopes (13 vs 14 WT), pI 6.40 (high solubility), GRAVY -0.149 (hydrophilic). Single disulfide bond. Zero aggregation hotspots. Ready for recombinant E. coli / Pichia expression.
One Molecule, Three Diseases
The MIF/CD74/NF-kB neuroinflammatory cascade drives pathology across multiple diseases. The core axis is identical — what changes is the effector arm: demyelination in MS, motor neuron death in ALS, immunosuppressive tumor microenvironment in brain metastasis. JFIN-0001 blocks CD74 at the source.
JFIN-0001 binds CD74 extracellular domain (Kd ~28.6 nM). Same protein, same epitope, three disease contexts.
Demyelination via MIF/CD74 in brain microglia
Effector Arm
NF-kB -> HES5/DNMT1 -> MBP/PLP1 suppression (demyelination)
JFIN-0001 at 100 nM (Kd 28.6 nM)
Comparative NF-kB reduction
Original indication. Patent OEPM P202530290.
Motor neuron death via MIF/CD74 in spinal cord microglia
Effector Arm
NF-kB -> TNFa/C3 -> complement-mediated motor neuron killing
JFIN-0001 at 100 nM (Kd 28.6 nM)
Comparative NF-kB reduction
Combination: Riluzole (orthogonal: glutamate + SLC1A2)
Joki et al. (2024) MIF elevation in ALS CSF.
Tumor-associated macrophage reprogramming via MIF/CD74
Effector Arm
NF-kB -> TNFa/IL6 -> immunosuppressive TME + antigen presentation block
JFIN-0001 at 100 nM (Kd 28.6 nM)
Comparative NF-kB reduction
Data package prepared for Dr. M. Valiente (CNIO).
All results are computational predictions from the NeuroInteractome Digital Twin platform. ODE-based Hill-function kinetics with calibrated disease-state parameters. Experimental validation pending. Patent OEPM P202530290 (April 2026).
Who We Are
JFInnova is a computational drug discovery company born from the conviction that AI can design better biologics, faster. We combine single-cell transcriptomics, foundation models, and multi-scale Digital Twins to create next-generation nanobody therapeutics for neuroinflammation.
Full-stack computational biologist. Designed the NeuroInteractome pipeline from scratch: from snRNA-seq analysis and scGPT fine-tuning, through de novo VHH nanobody design and molecular dynamics, to ODE-based Digital Twin simulations and ESM-2 directed evolution. Patent author (OEPM P202530290).
Tech Stack
Scientific Intelligence
Curated analysis of the discoveries, regulations, and industry movements shaping the future of neuroinflammation therapeutics — and how JFInnova is positioned at the convergence.
Researchers at the University of Zurich have published landmark single-nucleus RNA sequencing data demonstrating that CD74+ microglial density correlates directly with progressive disability accumulation in MS lesions.
Sanofi's strategic acquisitions of single-domain antibody platforms signal a major industry pivot toward VHH-based therapeutics, driven by their superior tissue penetration and manufacturing simplicity.
The SPRINT-MS trial (MediciNova) demonstrated modest neuroprotection with ibudilast, but its non-selective MIF inhibition and off-target PDE effects highlight the ceiling of small-molecule approaches in neuroinflammation.
The FDA's updated framework for AI/ML-designed therapeutic candidates introduces new transparency requirements for in silico design pipelines, including algorithm provenance documentation and computational reproducibility standards.
The CDTI has opened its 2026 NEOTEC call with a dedicated Health-AI track, allocating 20 million euros to support deep-tech startups developing AI-driven therapeutics and diagnostics on Spanish soil.
Recent publications on receptor-mediated transcytosis and single-domain antibody formats confirm that VHH nanobodies below 15 kDa achieve meaningful CNS penetration — a pharmacokinetic advantage inaccessible to conventional IgG antibodies.
Stay at the frontier
JFInnova continuously monitors the scientific landscape to ensure our platform and therapeutic candidates remain aligned with the latest evidence. For partnership or media inquiries, reach out through our Partnering section.
Partnering & Pipeline
JFIN-0001 is currently Pre-Clinical (In Silico). We are actively seeking academic or industry partners equipped for recombinant VHH expression and SPR/BLI binding assays.
Recombinant VHH expression (E. coli / Pichia) and SPR/BLI binding assays to advance JFIN-0001 toward in vitro validation.
JFIN-0001 available for out-licensing. Platform available for sponsored research in any neuroinflammatory indication.
Access to our Digital Twin platform for validation studies. Shared publication opportunities on novel targets.
NeuroInteractome platform available for clinical research groups. Custom target discovery and Digital Twin analysis for any inflammatory disease.
Full preclinical dossier, structural data, and competition analysis available under NDA. Contact fsalinas@jfinnova.es for investor deck.
Get in Touch
Interested in collaboration, licensing, or learning more about our platform? Send us a message.
info@jfinnova.es
Location
Madrid, Spain
Full preclinical dossier, structural data, and patent documentation available under NDA. We respond within 24 hours.