The Evidence Layer for Scientific AI

 Quick Explanation Copied

 Long Explanation

Core Scientific Track Record

Jacob E. Corn's publication record spans structural/computational biology (Rosetta3, RosettaScripts, computationally designed influenza binders), CRISPR methodology (asymmetric donor DNA HDR enhancement, CRISPRi/a library design, nanoparticle RNP delivery), translational editing (selection-free correction of the sickle mutation in human HSPCs), and current genome-editing chemistry. OpenAlex attributes 166 works and 16,114 citations with an h-index of 52 to his ORCID record, with citation intensity rising steadily since 2012 and peaking at ~1,600+/year in 2020–2026.

His recent first-class work includes a DarT2-nCas9 "append editor" that adds targeted DNA ADP-ribosylation as a distinct editing modality, rigorously dissected across nine genetic backgrounds in E. coli showing RecA dependence, with cross-kingdom outcomes in yeast, plants, and human cells Append editing genetic-dependence evidence. His group also built a systems-scale phosphosite-function predictor validated experimentally at four sites (PPM1G S201/S183, KHSRP S125/S274), with deposited code and BioStudies data Phosphosite function prediction preprint. Earlier lab contributions (modified gRNA/donor DNA platform enabling gDonor conjugates and fluorescent donor enrichment, ~8% HDR via polymer delivery) show the same method-driven, engineer-first orientation Modified gRNA/donor DNA eLife paper.

Citation Metrics: A Discrepancy to Flag

The supplied Semantic Scholar author record (h-index 2, 16 citations, 5 papers) conflicts with the OpenAlex record (h-index 52, >16,000 citations) and lists unrelated titles, indicating an author-disambiguation failure rather than genuine output. Judging Corn by that fragment would invert reality; the OpenAlex trajectory, top works (Rosetta3 ~1,945 citations; asymmetric donor DNA ~1,127; CRISPRi/a libraries ~896), and co-authorship with Doudna, Baker, Baltimore, and Church are mutually consistent and credible. A minor caveat: many landmark works are multi-author collaborations where Corn is middle/last author, so some citation mass reflects team science.

Rigor, Novelty, and Blind Spots

Evidence quality is strong across extracted records: controlled catalytic-dead variants, systematic repair-gene knockouts, multi-organism validation, and deposited data/code recur in his papers. Limitations flagged by the records themselves include motif constraints and DarT2 cytotoxicity for append editing, sparse phosphoproteomics coverage (~8,300 of ~189,576 sites per perturbation), and mostly cell-line models with limited in vivo validation Append editing limitations. Declared industry roles (Serac co-founder/SAB; SAB roles at Relation Therapeutics, Hornet Bio, Kano; collaborations with Allogene and Cimeio) are transparently disclosed and warrant awareness of commercial-incentive bias when evaluating translational claims Corn conflict-of-interest disclosure. Overall, the evidence supports a top-tier, method-defining scientist whose weakest exposed flank is translational/clinical validation depth rather than conceptual rigor.

 BGPT Author Review

Scientific Quality

The OpenAlex record (166 works, ~16,114 citations, h-index 52, sustained >1,400 citations/year since 2020) plus first-class methodological outputs (append editing with clean genetic-dependence controls; phosphosite ML with experimental validation; modified gRNA/donor DNA platform) justify a top-tier score. The Semantic Scholar fragment (h-index 2) is an obvious identity-split artifact and was discounted. Minor deductions: heavy reliance on collaboration-derived citation mass, cell-line-dominant validation, and motif/toxicity constraints in the newest technology.

Communication Quality

Papers consistently pair mechanisms with usable tools (web apps, deposited code, source data), aiding dissemination; multi-organism presentation in append editing is clear and logically ordered. Deductions because communication quality of talks/figures could not be directly assessed from supplied extracts, and some work depends on dense multi-omics methods that raise reader burden.

Author Novelty

Repeatedly opens new modalities rather than incrementally extending: ADP-ribosylation append editing as a chemically distinct editing axis, gRNA-donor conjugates, phosphosite-function prediction for the dark phosphoproteome. Not a perfect 10 because base/prime editing context and CRISPR-delivery improvements build on established frameworks.

Scientific Rigor

Extracted records show catalytic-dead controls, knockout panels across nine repair genes, triplicate biological replicates, deposited sequencing data, and explicit limitation sections (motif constraints, TARG1 dependence, off-target ADPr risk). Deduction for in-vivo validation gaps and undisclosed replicate counts in the 2017 eLife platform paper.