Assess an author's data and outputs

Rapid verdict — Qihong Yan (author)

Metrics: ~27 papers, ~600 citations, h-index ~15 (author-supplied); OpenAlex shows a related profile with h-index 18 and ~857 citations (possible multiple name-variants). Publication portfolio concentrates on SARS-CoV-2 and antibody/BCR immunology with several experimental and computational works, moderate citation impact, and clear strengths in B-cell repertoire and antibody characterization. Key recent methodological advance to note: InterAb (Jan 21, 2026) claims high predictive performance for antibody specificity/affinity using all-atom modeling + pre-trained language models (SPE-Test AUC 0.9459; AFF-Test Pearson 0.6877) and wet-lab validation (BLI + pseudovirus assays), indicating competence with computational–experimental integration InterAb 2026 Computational+Experimental Paper

Author Review: Qihong Yan — Visual, Evidence-based Appraisal

Notes: Author-supplied metrics indicate modest productivity (27 papers) with moderate citation impact (~600 citations, h≈15). OpenAlex returns a nearby profile (likely same researcher or variant) with higher counts (43 works, 857 citations, h≈18) — this discrepancy suggests name disambiguation issues common in bibliometrics (multiple ORCID/name variants) and motivates careful identity matching before bibliometric conclusions.

Evidence summary (claims supported by provided papers)

• Concentration on antibody biology, B-cell receptor (BCR) repertoires, and neutralizing monoclonal antibodies against SARS-CoV-2 and related viruses (multiple provided titles show recurring motifs: "antibody repertoire", "neutralizing antibodies", "SARS-CoV-2", "spike", "RBD").
• Combines experimental immunology (BLI kinetics, neutralization assays, animal protection studies for Zika) and computational approaches (recent InterAb paper integrating all-atom modeling with language models) indicating an interdisciplinary skill set.
• Moderate bibliometric impact (h ≈ 15–18; total citations several hundred), consistent with an active mid-career researcher producing influential niche work but not (yet) broad field-leading citation metrics.

Key methodological contribution to evaluate

InterAb (Jan 21, 2026) — computational+experimental pipeline for predicting antibody specificity and affinity and optimizing breadth. Major claims: high SPE-Test AUC (0.9459), AFF-Test Pearson ~0.6877, broadening optimized antibodies with experimental BLI and pseudovirus neutralization validation. Strength: combines all-atom interface modeling (AtomInter) and large pre-trained antibody language models; includes wet-lab validation for optimized sequences. Main limitations acknowledged: reliance on predicted structures (ESMFold/Chai-1 when experimental data missing), limited publicly-available datasets and code, possible dataset composition biases, and potential generalization gaps beyond SARS-CoV-2/Influenza A InterAb 2026 Computational+Experimental Paper

Strengths (evidence-backed)

• Domain focus: repeated experimental studies on SARS-CoV-2 antibodies, BCR repertoire dynamics, and neutralization — indicates domain expertise and coherent research program (multiple paper titles provided).
• Integration of computational modeling and experimental validation (InterAb + BLI/neutralization) — an important trend in modern antibody engineering and an indicator of methodological sophistication InterAb 2026 Comp–Exp.
• Experimental breadth: papers include repertoire sequencing, monoclonal antibody isolation, and animal-model protection studies (Zika) — showing competence across wet-lab immunology methods.

Concerns, blindspots, and potential biases

• Bibliometric ambiguity: author name disambiguation (OpenAlex returns multiple Qihong Yan variants) complicates accurate counting of works/citations — always verify ORCID and institutional affiliation before conclusions.
• Reproducibility risks for computational work: InterAb notes limited public code/data release and heavy reliance on large pre-trained corpora and predicted structures (ESMFold/Chai-1). Without full open release and independent replication, performance claims are provisional InterAb limitations.
• Possible dataset bias: curated datasets (SPE7626, AFF1735, SKE438) may over-represent certain antibody families or antigens, inflating reported generalization; independent external validation sets are necessary.
• Generalizability: most work centers on SARS-CoV-2 and Influenza A; claims of broad-spectrum utility should be tested on non-viral protein antigens and in diverse experimental systems.

Concrete recommendations for the author (to strengthen scientific standing)

1. Publish or release code, model weights, and standardized dataset splits (SPE7626, AFF1735, SKE438) to enable independent reproduction and benchmarking.
2. Provide thorough ablation studies quantifying performance dependency on predicted vs experimental structures (e.g., run InterAb on only experimentally determined complexes vs only predicted complexes and report delta metrics).
3. Perform independent external validations on antigens outside SARS-CoV-2 / Influenza A to demonstrate generality (e.g., bacterial toxins or tumor-associated antigens).
4. Resolve author disambiguation (provide ORCID and primary institution consistently) to consolidate bibliometric records and improve discoverability.

How to falsify the main InterAb claim (explicit testable criteria)

Failure modes that would refute InterAb’s central claims include: (1) when evaluated on a truly independent unseen-antigen benchmark, specificity/affinity predictions fall to baseline or below competing methods; (2) optimized sequences do not show improved binding or neutralization in blinded wet-lab tests; (3) performance drops dramatically when only predicted structures (no experimental structures) are used — all measurable and falsifiable outcomes InterAb falsification criteria.

Bottom-line (evidence-weighted): Qihong Yan appears to be an active researcher focused on antibody immunology and computational antibody engineering with a moderate citation footprint and some recent high-ambition methodological work (InterAb) that bridges computation and experiment. The scientific promise is high if reproducibility, public code/data, and external validations are provided; current limitations are typical for cutting-edge computational–experimental studies (dependence on predicted structures, dataset curation, and limited data/code availability) and should be addressed to turn promise into durable field impact InterAb cautionary note.