Ketchem Biotherapeutics Consulting LLC

ketchemconsulting.com ablead@ketchemconsulting.com

AbLead

Developability by Design

Antibody sequence and structure analysis to evaluate, rank order, and engineer biotherapeutics.
Identify and engineer stable, manufacturable candidates before you invest in the lab.

The Cost of Selecting the Wrong Lead

Pursuing a poorly behaving antibody is one of the most expensive mistakes in drug discovery.
Sequence and structural analysis and optimization minimizes this risk.

$11K-$16K Average daily out-of-pocket cost for the early development of a biotherapeutic lead candidate .
60% Phase 2 antibody candidates with developability issues .
$800K Lost value per day of delay due to poor developability .
"The cost of engineering an antibody at the discovery stage is significantly lower than the cost of optimizing process and formulation at the development stage ."
"Application of rank ordering and stability optimization is a proven method for antibody therapeutic design ."
Randal R. Ketchem, Ph.D.

Biotherapeutic optimization is well worth the initial investment of time and resources

Biophysical Properties

Stability engineering greatly improved overall biophysical properties ...

Stable Cell Titer

... and more than doubled stable cell titer .

Serum Half Life

Developability optimization markedly improved serum half-life .

Empirical validation proves the case for early stage developability optimization.

Capabilities

Explore Detailed Platform Documentation & User Guide →
01

Rank Order Analysis & Lead Selection

Evaluate antibody sequences across a wide range of developability attributes and predictive machine learning models to rank order and select lead candidates.

  • KBC Score: Aggregate developability score weighted by severity allows for cross-trait rank ordering and targeted lead selection.
  • Lead Panel Selection: Rank order and select candidates based on customizable scoring weights, thresholds, and filtering criteria.
  • Predicted Biophysical Properties: Multi-modal machine learning forecasting 10 quantitative developability endpoints calibrated on clinical-stage therapeutics .
  • RPEMHC Immunogenicity: Deep learning 15-mer sliding window MHC Class II binding prediction across an expanded 20-allele global panel with population-weighted ADA risk scoring .
  • Covariance Analysis: Statistical OMES and Pairwise Conservation Score (PCS) analysis to detect and repair broken structural contacts and co-evolving residue constraints .
  • Automated Fv Profiling: Granular multi-attribute analysis including:
    • Germline Mapping: Closest V/J-Gene identity and origin mapping across chains.
    • Nearest Neighbor: Identification of nearest sequence neighbors within antibody sequence space.
    • CDR Lengths: Evaluate CDR loop lengths and total CDR sum.
    • Structural Stability: AbLang, AbLang2, and IgBert machine learning likelihood scores predicting conformational stability .
    • Salt Bridge Detection: Identification of disrupted HC-CDR3 salt bridges that impact CDR stability and efficacy .
    • Surface Properties: Surface Hydrophobicity (SPH), Positive/Negative Charge Patches (SPP/SPN), and Fv Charge Dipole (SPCD) calculation .
    • Cysteine Analysis: Detection of unpaired, missing, or non-canonical cysteines to ensure stable disulfide pairing.
    • Potential PTMs: Deamidation, isomerization, oxidation (Met/Trp), N-glycosylation, hydrolysis, and fragmentation scanning .
    • Humanness (OASign): Log-likelihood scoring against 130M+ human antibody sequences.
    • Isoelectric Point (pI): Theoretical (Bjellqvist) and empirical structure-based (PROPKA) pI evaluation to optimize formulation and avoid serum instability.
02

In-Silico Engineering & Optimization

Design, evaluate, and build optimization variants, combinatorial libraries, and humanized candidates directly within the workspace.

  • First Pass Optimize: Automated 1-click optimization engine scanning and repairing developability liabilities and language model penalties with structural solvent accessibility gating.
  • Humanization & Bulk Humanize: Germline template or Sapiens-based humanization with controlled site exclusions, plus automated high-throughput bulk humanization across lead panels.
  • Combinatorial Variant Generation: Multi-set design workspace with mutual exclusivity and linked pairing rules (`Pair Single` / `Pair Together`), Min/Max mutation filters, and Deep Mutational Scanning (DMS) permissibility integration.
  • Isoelectric Point Engineering: Shift candidate pI toward basic or acidic ranges with solvent-accessible charge redesign while monitoring humanness and stability.
  • Surface Properties & Mutagenesis: Visualize and redesign residues contributing to hydrophobic patches, charge clusters, pH switches, or binding interfaces.
  • Batch & Multispecific Assembler: Assemble full-length IgG, bispecifics, CrossMabs, scFvs, and complex heterodimeric formats with pre-configured linkers and Knobs-into-Holes domains.
  • Fv Pair Extraction & Junction Repair: Automatically dissect VH/VL variable domains from multispecifics and correct constant light chain V-J junction frame defects.
  • 3D Structural Modeling: Generate high-accuracy 3D structural Fv models directly in-browser using ABodyBuilder2 .
  • Observations Log: Record and search IMGT-positioned residue observations and experimental history across projects to guide engineering choices.
03

Visual Analysis, Discovery & Reporting

Interactive dashboards and discovery tools for multi-parameter exploration, phylogenetic diversity, structural mapping, and compliance.

  • Interactive Multi-Parameter Explorer (Plot Results): 2D Scatter Plots, Box Plots (with violin contours), Binned Histograms, and Bar Rankings with automated 1-vs-all Best Fits correlation discovery and multi-variable combinatorial regression.
  • Clading & Lineage Evolution: Phylogenetic clustering with Somatic Hypermutation (SHM) divergence from unmutated germline precursors (UCA), ancestral vs. private branch developability liability glyphs, and circular sector layouts.
  • Numbering Schemes & CDR Regions: Aho, IMGT, Martin, and Kabat numbering assignments with Aho, IMGT, Martin, Kabat, and North CDR region detection .
  • Sequence Alignments: Numbering system-based alignments with consensus calculation, germline cluster shading, and direct residue comparisons.
  • 3D Structure & Structural Liabilities: Interactive in-browser Molstar 3D structure viewer with per-residue liability mapping, solvent accessibility, and PyMOL/MOE SVL export capability.
  • Humanness & Residue Analytics: Multi-method evaluation against human germlines with per-residue OASign, AbLang Diff, Positional Frequency Analysis (PFA), and 2D MHC-II peptide presentation matrix heatmaps.
  • Mutation Grid & Physical Properties: Side-by-side mutational difference matrices between parent and variant clones alongside panel-wide physical property dashboards.
  • Clinical Comparison: Benchmark candidate developability profiles against the Jain clinical-stage therapeutic antibody landscape .
  • PLAbDab Patent & Literature Search: Perform internal similarity searches against the PLAbDab patent and literature antibody database.
  • WIPO ST.26 Patent Sequence Listing: 1-click generation of patent-compliant WIPO ST.26 XML sequence listings for global patent filings (USPTO, EPO, WIPO).
  • Comprehensive Analysis Reports: Export structured Word (`.docx`) audit reports and formatted, color-coded multi-sheet Excel workbooks.

References