Review 1 of “A Multi-Source Profile of TAAR3P: trace amine associated receptor 3, pseudogene as an IDG Orphan G Protein-Coupled Receptor” (v4)

Summary and recommendation

Recommendation: major revision. This review concerns version 4.

The manuscript assembles a resource-based profile of human TAAR3P, reporting sparse GeneRIF coverage, two GO biological-process annotations, no returned GSFM or regulatory results, a single ARCHS4 expression row, and 403 Harmonizome associations. A transparent inventory of available and missing annotations could be useful. The explicit distinction between Enrichr annotation lookup and enrichment testing is a strength. However, the conclusions about a druggable GPCR exceed the evidence, and unresolved identifier mapping, missing-data handling, and provenance problems prevent the reported profile from serving as a reliable baseline.

I reviewed the complete version-4 Markdown and checked external annotation and resource references. I did not rerun the authors' pipeline or verify the attached PDF, which was inaccessible through my retrieval tools. The observations about tables and captions below refer to the Markdown.

Major issues and requested changes

1. Correct the biological premise and therapeutic framing. NCBI Gene 9288 identifies human TAAR3P as a pseudogene, with the noncoding RefSeq RNA NR_028511.1. The manuscript supplies no evidence of a functional human receptor protein. Predicted receptor annotations do not establish translation or receptor activity. Revise the title, abstract, and discussion accordingly; distinguish human-locus evidence from protein-family or ortholog-derived annotations. Replace routine recommendations for biochemical receptor assays and structure-based drug design with a staged plan to establish locus-specific transcription and, if claimed, translation and activity. Pseudogene status does not rule out biological function, but it changes which hypotheses are supported.

2. Demonstrate identifier resolution and IDG membership. The IDG classification field is empty in the summary, and the abstract contains a missing category. Low GeneRIF coverage cannot by itself establish IDG category membership. Provide the exact classification API response and its mapping to Entrez 9288, species, and symbol. Also audit aliases: the Harmonizome TAAR3 page maps to Entrez 9288, lists TAAR3P as a synonym, and reports 396 associations across 30 datasets in the retrieved page, whereas the manuscript reports 403 associations from one dataset. This discrepancy does not establish that either count is wrong; it could reflect releases, endpoints, or filters. Supply a per-resource identifier crosswalk and archived responses, explain the discrepancy, and verify that the search did not omit relevant records or merge distinct species/loci.

3. Resolve the expression result before drawing tissue conclusions. The tissue table contains only “System,” with median 0.00 and z-score −1.11. That is not an interpretable tissue profile without its source metadata. The reported z-score is consistent with the stated formula, (0 − 4.336)/3.921, but the arithmetic cannot distinguish a measured zero from a missing value converted to zero. Provide the raw response, actual tissue labels, sample counts, expression units/transformation, quantification and identifier mapping, and missing-value rules. Explain the origin and suitability of the random-gene baseline. Remove claims of tissue-specific enrichment unless supported by valid tissue observations; this descriptive standardization is not a significance test.

4. Reconcile Harmonizome reporting and clarify what the scores mean. The abstract says multiple datasets, the results name one LINCS L1000 chemical-perturbation dataset, and the figure caption names JASPAR TF predictions. Identify the dataset underlying each table and image and regenerate inconsistent captions. Deposit all 403 rows with dataset identifiers and score definitions; specify whether ranking uses signed scores or absolute magnitude. For LINCS results, document whether TAAR3P expression is directly measured or inferred, and provide available perturbation context and compound identities. Expression responses to compounds do not establish compound binding, receptor agonism, or therapeutic efficacy. Explain dataset-specific standardization and avoid presenting scores as calibrated statistical significance without an appropriate analysis.

5. Make null results and literature coverage reproducible. Include code, query dates, API/library versions, request parameters, raw responses, and a status table distinguishing successful empty results from unsupported identifiers, failed requests, and filtered invalid values. This is especially relevant after revisions removed NaN similarity outputs. Provide the PMID behind the single GeneRIF record. Describe that count as coverage of the queried GeneRIF resource rather than a comprehensive PubMed census. Keep negative claims scoped to the source and query; absent returned regulatory or model records do not establish biological absence.

Additional corrections

6. Repair citations and template remnants. References 3 and 6 do not match the resource papers I verified: Geneshot is Lachmann et al., 2019, DOI 10.1093/nar/gkz393, and the original Harmonizome paper is Rouillard et al., 2016, DOI 10.1093/database/baw100. Audit all six references and add stable identifiers. Correct the description of CFDE: its NIH program overview describes reuse and integration of Common Fund data broadly, rather than a fibrosis-specific consortium. Remove the empty literature-network heading, the duplicated family wording, and statements promising analyses absent from the results. Check that the attached PDF corresponds to version 4 and the same human locus.

A revised descriptive resource profile could be worthwhile without new wet-lab experiments, provided it resolves provenance and mapping, treats missingness explicitly, and limits conclusions to supported annotations. Retaining claims of a functional, druggable human GPCR would require additional evidence.