Our study began with a contradiction.
Comparative genomic analyses suggested that GPRC6A, a gene involved in nutrient sensing and metabolic regulation, had accumulated potentially disruptive changes in cattle and related species. Yet experimental studies had reported GPRC6A-associated activity in bovine mammary epithelial cells, where the gene had been linked to pathways involved in milk-fat synthesis.
These two bodies of evidence seemed difficult to reconcile. If the gene was genuinely disrupted, how could evidence of transcription, translation, or biological activity persist? Conversely, if functional activity remained, was it appropriate to describe the gene simply as “lost”?
Rather than assuming that one side of the literature must be wrong, we treated the discrepancy as a starting point.
Looking beyond a binary definition of gene loss
Gene annotation often relies heavily on coding-sequence integrity. Premature stop codons, frameshifts, insertions, deletions, and disruptions of an open reading frame are commonly interpreted as signs that a gene has become nonfunctional.
This approach is extremely useful for comparing thousands of genes across many genomes. However, biological systems do not always fit neatly into binary categories. A sequence may appear disrupted while still producing a shortened transcript, a small peptide, an alternatively initiated protein, or another biologically active product.
We therefore asked a broader question:
Could GPRC6A retain some form of biological activity even if its conventional full-length coding sequence had been disrupted?
Our objective was not to prove one particular mechanism. Instead, we wanted to organize the available genomic, transcriptomic, proteomic, and experimental evidence into a set of plausible and testable explanations. The resulting article evaluates five such scenarios.
Five possible routes to retained activity
The first possibility is the production of a truncated protein or micropeptide. A disrupted open reading frame does not necessarily prevent all translation. Shorter open reading frames may remain intact, and some can generate stable products with biological functions of their own.
This possibility became especially interesting when we examined publicly available proteomic datasets. A peptide mapping to the bovine GPRC6A region had been detected in bull-sperm proteomic data. The peptide does not establish the presence of a complete, canonical receptor, but it provides evidence that at least part of the locus may be translated.
The second possibility involves alternative transcription initiation. A transcript does not always have to begin at the conventional start site. Internal promoters or downstream transcription start sites can produce shorter RNA molecules that bypass disruptive mutations located nearer the beginning of a gene.
To explore this possibility, we examined predicted and experimentally derived transcription start sites across the bovine GPRC6A region. Several candidate sites occurred downstream of the conventional start position, making internal transcription a plausible hypothesis that can be tested experimentally.
The third scenario is translational recoding. Ribosomes occasionally use mechanisms such as stop-codon readthrough or programmed frameshifting to continue translation across sequence features that would normally interrupt a protein. We identified sequence motifs and nearby structural features that are compatible with such mechanisms. Their presence does not demonstrate that recoding occurs, but it provides concrete regions for future experimental investigation.
The fourth possibility is RNA-level processing or repair. Editing, unusual splicing events, or the removal of disruptive sequences from an RNA transcript could, in principle, alter the relationship between the genomic sequence and the final translated product. Repetitive elements within the GPRC6A region may form RNA structures relevant to such processes.
These mechanisms remain speculative at this locus. Nevertheless, related forms of RNA processing have been documented elsewhere, making them biologically plausible rather than purely hypothetical.
The fifth and least-supported possibility is the existence of an unresolved genomic copy. Complex or repetitive regions can remain incomplete even in high-quality genome assemblies. An additional intact or partially intact copy could therefore have escaped current annotation. We found no direct evidence for such a copy, so we regard this explanation cautiously. Future haplotype-resolved and long-read assemblies could test it more rigorously.
The central schematic in Figure 1 of the article brings these possibilities together, showing how different transcriptional, translational, RNA-level, and assembly-related mechanisms might allow activity to persist despite apparent sequence disruption.
Working with evidence that did not agree neatly
One of the most challenging parts of the project was resisting the temptation to force the evidence into a single narrative.
The transcriptomic data did not show strong, uniform expression across the entire bovine GPRC6A locus. In the public RNA-sequencing datasets we examined, cattle generally showed lower transcript abundance than humans and pigs. No individual bovine sample showed clear expression across all six examined exons.
At the same time, several observations prevented us from treating the locus as completely silent. These included partial transcriptional signals, downstream transcription start sites, short open reading frames, and the peptide detected in a bovine proteomic dataset.
Each observation was incomplete on its own. Together, however, they suggested that the usual categories of “present” and “lost” might not adequately describe the locus.
This led us to treat the study as hypothesis-generating, rather than as a definitive demonstration of how GPRC6A activity is maintained.
The idea of a “peripatetic molecular fossil”
We introduced the term “peripatetic molecular fossil” to describe a locus that appears displaced from the category in which sequence-based annotation would normally place it.
The phrase was inspired by historical palaeontological puzzles in which fossils appeared outside their expected geographical or stratigraphic settings. In our case, the unexpected object is a gene that appears eroded at the sequence level while retaining evidence suggestive of transcription, translation, or biological activity.
We proposed three practical features for recognizing such a case:
- The locus contains sequence changes normally associated with gene loss.
- Independent evidence suggests activity in at least one biological context.
- The molecular source of that activity remains unresolved.
The term is not intended to create another rigid annotation category. Rather, it highlights cases that deserve further investigation before being reduced to a simple present-or-absent label.
What the study does not establish
An important part of the project was defining its limits.
We did not generate new wet-laboratory data. We have not demonstrated that any one of the five proposed mechanisms operates in bovine mammary tissue. The proteomic peptide requires further biochemical validation, and the predicted transcriptional and recoding mechanisms remain provisional.
The article therefore does not claim that canonical full-length GPRC6A has been restored or that its exact molecular form in cattle has been identified.
Instead, it argues that the available evidence supports keeping the question open.
Where the research could go next
Several experiments could help distinguish among the proposed explanations:
- tissue-specific, full-length transcript sequencing;
- ribosome profiling to identify actively translated regions;
- targeted mass spectrometry of candidate peptides;
- biochemical characterization of truncated products;
- long-read and haplotype-resolved reconstruction of the locus;
- functional assays testing individual predicted transcripts or open reading frames.
These experiments would clarify whether the biological activity arises from a truncated product, an alternative transcript, translational recoding, RNA-level processing, an unresolved genomic copy, or another mechanism not yet considered.
What we learned
The most important lesson from this project was methodological.
Sequence comparison is extraordinarily powerful, but sequence erosion does not always provide the final word on biological activity. Functional evidence can also be incomplete or context-dependent. The strongest interpretation often emerges only when genomic, transcriptomic, proteomic, and experimental observations are examined together.
Our study does not close the GPRC6A question. It narrows the problem by organizing the plausible explanations, and identifies experiments that could resolve it.
That, for us, is the value of examining an apparent contradiction. It turns a difficult annotation into a testable research programme.
Competing interests
We are authors of the present article and co-authors of the earlier comparative genomic study that helped motivate it. This post reflects our interpretation of the published and publicly available evidence. We declare no other competing interests relevant to this post.
AI use declaration
Generative AI was used to assist with language editing and organization. The authors reviewed and verified the scientific content and take responsibility for the final text.