BGPT: Author Review: Ye Yang

Explore by Goal

Quick Explanation Copied

Ye Yang — Scientific strength check

Based on the papers you provided (mostly mechanistic and assay-heavy), Ye Yang’s work shows cross-domain biological modeling (plant single-cell+spatial metabolomics, complement/antibody pharmacology, immune-cell differentiation, and proteome-wide ligand binding), but several entries also exhibit typical translation/reproducibility risks (in vitro or single-model reliance; partial sampling; and inference-heavy interaction networks).

Key example strengths: mechanistic coupling + multi-assay validation in Panax single-cell/spatial metabolomics ().

Long Explanation

Author Review: Ye Yang

I’m going to be strict about evidence: your prompt provides (i) author bibliometrics (h-index/citations/paper count) but no source-IDs (DOIs) to cite for those metrics, and (ii) a set of paper-level research extracts with DOIs. Therefore, the scientific critique below is grounded only in the paper extracts you supplied, which span multiple biological subfields.

Evidence map (what is actually “strong” here)

Multi-modal mechanistic coupling (cell-fate trajectories + spatial metabolite maps + reporter/EMSA validation) in Panax root tips work ().
Orthogonal biophysical + functional assays for complement lectin-pathway targeting (binding kinetics by SPR, lectin-pathway functional readout by C4 deposition, and PK/PD in non-human primates) in the anti-MASP-2 antibody paper ().
Mechanism-first receptor–ligand biology linking a gut lipid metabolite to a nuclear receptor (SPR binding + reporter activation + cell differentiation + infection protection with dependence on genetic knockout) in the Nur77/12R-HETE study ().

Figure 1 — Panax scRNA-seq input vs QC-kept cells

The extract reports PN: 5761/6036 high-confidence/initial; PG: 7026/7424; PQ: 11638/12837; totaling 24,425 cells used in the integrated analysis ().

Figure 2 — anti-MASP-2 selectivity: SPR KD across species (from extract)

The provided extract reports SHR-2010 KD ≈ 3.49×10⁻¹⁰ M (human), ≈ 1.38×10⁻¹⁰ M (rhesus), and much weaker binding ≈ 3.81×10⁻⁸ M (mouse) ().

Critical appraisal (strengths vs failure modes)

1) Mechanistic depth vs network inference

The Panax and Nur77-style papers are mechanistically ambitious, but a common scientific risk is that interaction networks can become correlation-over-causation scaffolds. In the Panax extract, ligand–receptor links are described as derived from PlantPhoneDB and cross-species inference, which the extract explicitly flags as needing caution ().

A stronger pattern appears where assays tightly close the loop: Nur77 work ties (i) ligand binding (SPR), (ii) receptor activation (reporters), (iii) cell subset differentiation (ILC3 phenotyping), and (iv) pathogen outcome with genetic dependence (Nur77 knockout) ().

2) Cross-species translation is handled (but not eliminated)

The complement antibody extract demonstrates explicit awareness of species-binding mismatch and uses a surrogate to model murine PD. That’s good experimental hygiene, but it also means mouse efficacy is indirectly tethered to the human-target biology ().

3) Reproducibility and transparency: what we can/can’t verify

For several provided entries, the extract mentions public deposition (e.g., Panax scRNA-seq raw data in NCBI Bioproject accessions) (), which supports independent re-analysis.

But other extracts explicitly note limited transparency (e.g., data not publicly posted; code unavailable; “request from corresponding author”), which creates an avoidable verification gap. The complement antibody extract, for instance, states raw data are not publicly posted but may be requested ().

4) Biological blind spots to actively look for (based on the extracts)

Protoplasting / dissociation effects (Panax): can shift stress-response programs and thus bias fate trajectories and ligand expression patterns ().
Inference-based ligand–receptor specificity: cross-species homologs may mis-assign signaling partners even when spatial expression is compelling ().
System-level confounds in in vivo receptor/ligand work (Nur77): pharmacological inhibitors can have off-targets; developmental timing can limit translation ().

Overall scientific strength judgment (from provided extracts)

Strength signal: Across the provided DOIs, the author appears to work in a “mechanism closure” style: binding/functional assays, multi-omics readouts, and dependence testing via genetics or pathway perturbations. That pattern is strongest in the Nur77/12R-HETE chain and in the anti-MASP-2 antibody’s SPR → complement functional readout → PK/PD → in vivo surrogate efficacy design ( ).

Weakness signal: Several extracts still carry “standard biomedical” risks: protoplasting/dissociation artifacts, reliance on database-inferred ligand–receptor edges, limited sample sizes (or only a subset for long-read sequencing in surveillance-type work—though that’s outside the figures shown here), and incomplete public data/compute transparency for full independent replication ( ).

What would disprove or substantially change confidence?

Panax ginsenoside pathway model: direct Panax functional tests that break the proposed TF→enzyme/gene effects (e.g., MYB78/MYB2/IAA29 dependence) without changing upstream differentiation could falsify parts of the regulatory coupling ().
Nur77/12R-HETE immune differentiation: if Impdh1 is not necessary, or if Notch dependence fails under alternative genetic context, the mechanistic pathway would weaken ().
Complement antibody translation: failure to reproduce sustained lectin-pathway inhibition (C4 deposition readout) with similar PK/PD in additional primate cohorts would undermine confidence in the “long-acting” thesis ().

Note: I did not include any treatment recommendations. This review is limited to the mechanistic evidence in the provided extracts.

Feedback:

Updated: March 30, 2026