ImmunoPrecise Antibodies Ltd. (NASDAQ: IPA) (“IPA” or the “Company”), a Bio-Native AI pioneer operating where TechBio meets true biological intelligence, today announced results of a newly expanded study demonstrating that its LENSai™ Immunogenicity Screening can reliably predict Anti-Drug-Antibody (ADA) risk for therapeutic proteins before they enter animal studies or human trials.
“ADA-related failures still derail up to 40 percent of biologics in late development, costing companies billions of dollars,” said Dr. Jennifer Bath, President & CEO, IPA. “With LENSai, we can now triage candidates against clinical data in hours, not months, giving drug-developers a fast, objective way to de-risk programs early and focus resources on the safest molecules.”
Late‑stage ADA failures can wipe out $1–2 billion in projected revenue for a single biologic and push launch timelines back 12‑18 months. Yet many programs still lean on time‑intensive lab assays or first‑generation in‑silico screens that look only at peptide‑to‑MHC II binding across a few dozen HLA alleles—leaving large swaths of immune diversity untested and immunogenic “self/non‑self” checks largely unaddressed.
LENSai’s HYFT‑powered,alignment‑free immunogenicity screening compresses that entire workflow into one overnight run. It evaluates nearly 900 HLA variants and performs a whole‑proteome “humanness” scan at residue resolution, instantly flagging hot‑spots that legacy tools miss. By revealing design‑level fixes before expensive animal studies or repeat GMP production, the platform can shorten pre‑clinical cycles, cut material costs, and de‑risk downstream development—delivering a speed‑and‑breadth advantage unattainable with traditional methods.
Study Highlights
- 217 marketed and clinical-stage antibodies analyzed – the largest public ADA dataset to predict immunogenicity risk
- Single composite score tracks clinical reality – enables reliable ADA incidence risk classification; a score ≥ 54 flags high-risk candidates (> 30 % ADA in patients). The discriminative capability is very powerful as indicated by an AUC=0.92.
- Alignment-free, HYFT-powered engine – the proprietary screening algorithm combines state-of-the-art MHC II binding with IPA’s proprietary HYFT® patterns, enabling whole-proteome humanness assessment without multiple-sequence alignments.
- Detailed amino acid and epitope-level immunogenicity “self” scanning – quickly pinpoints immunogenic hot-spots and suggests sequence edits before expensive wet-lab assays.
Read the full case study: [link]