Focused OpenScientist investigations of fly function hypotheses

Four unresolved predictions have a concrete route to further mechanistic investigation: mitochondrial-carrier substrate specificity, crotonase-family reaction assignment, catalytic competence of a short glycosyltransferase isoform, and cross-species ligand recognition. These cases were selected from the 41-gene fly cohort because comparative or structural analysis could add decisive evidence beyond the completed literature reviews.

Fly cohort and reviews · Selection manifest · Job status snapshot · OpenScientist jobs

Selected hypotheses

Target	Exact claim being investigated	Why a further investigation is useful
Dic4, Q9VVS1	Uptake of thiamine pyrophosphate into mitochondria	Mitochondrial-carrier architecture does not establish substrate specificity. Compare experimentally characterized carrier subfamilies, specificity residues and the actual substrate panel in Dic4 transport studies. A negative assay with one substrate does not exclude another.
CG5611, Q9VB17	An enzymatic contribution to fatty-acid oxidation	The crotonase superfamily supports many different reactions. Subfamily placement, catalytic geometry and characterized homologs could identify the reaction and distinguish a fatty-acid pathway role from competing metabolic functions. The original tentative wording is preserved.
ttv, D5SHU8	Glycosyltransferase activity of the 299-residue ttv-PC product	The short native product retains a catalytic region. Domain folding, substrate recognition and partner requirements matter more than length alone. The investigation separates intrinsic catalytic competence from targeting and participation in glycan synthesis in vivo.
TyrRS, Q9VV60	Resveratrol binding	A mammalian ligand-bound structure provides a testable pocket-level comparison. Conserved tyrosyl-tRNA synthesis does not establish the additional ligand interaction. Biochemical binding, downstream signaling and suitability as a GO annotation are separate questions.

All four current prediction assessments are UNC. That status is curator context and is withheld from the submitted prompts. For TyrRS, the emitted GO:1905594 is obsolete; the investigation tests its biochemical content without treating binding evidence as a reason to restore an obsolete term.

Investigation briefs

Each brief supplies the exact frozen sequence, its SHA-256 checksum, accession and FlyBase identity lead. The current sequence is not asserted to be the original prediction-time input. The provider receives a neutral hypothesis and bounded questions, with the existing review verdict and local analyses held back. Primary papers are leads to inspect, not preassigned conclusions.

Selection boundaries

The resolved protease/pseudoenzyme cases, explicit spider context in Lcp3, and routine conserved functions do not receive another investigation. Their decisive evidence is already available. Ank2's residual exocytosis claim is deferred because isoform-specific perturbation or interaction evidence may remain limiting. CG5565's proposed weak in-vitro side activity is also deferred: a different preferred physiological substrate cannot refute such activity, and a controlled assay is more likely to settle it than another fold comparison.

Interpreting the reports

The requested outputs include executed methods, actual results, sequence identities, primary citations and explicit limitations. Docking scores alone do not establish binding, a shared fold does not establish substrate specificity, and a conserved catalytic residue does not establish an active isolated domain. Conversely, absence of a target-specific experiment does not invalidate a well-supported conserved-family inference.

Returned reports will be compared with the held-out analyses and inspected source by source. Conclusions about other isoforms remain separate from conclusions about the submitted protein. The reports can support, refute, narrow or leave the hypotheses unresolved; the existing reviews are not changed merely because a job was submitted.

Execution and provenance

The four prepared commands passed dry-run checks. Each uses three OpenScientist iterations, a 7,200-second provider limit and an 8,100-second wrapper limit. The collector records accepted job IDs, upstream status and downloaded report paths. It refuses an existing local run-status file or matching remote hypothesis to prevent duplicate submissions. The static job-status link is a saved snapshot; the collector updates its repository source while jobs run.

Accepted jobs

All four investigations were accepted on 9 September 2026 UTC. Reports are pending; submission does not change the biological assessments. Submission verification confirms that each upstream job contains the prepared prompt and exact frozen sequence.

Target	OpenScientist job ID
Dic4	`cc51f444-45ab-4933-b480-688ac83406d5`
CG5611	`5f3a68b0-8473-4aca-b934-362c89e625b0`
ttv	`939966e3-12ee-4a0a-8311-48cd73058c14`
TyrRS	`30831c8b-4d81-4e64-a72f-8a3a97a9605d`