TAF14
S. cerevisiae
TATA-binding protein-associated factor TAF14
What the model saw
Gene: TAF14 Organism: Saccharomyces cerevisiae Known functions: TATA-binding protein-associated factor TAF14. DNA binding. RNA polymerase II general transcription initiation factor activity. histone binding. protein binding.
View unredacted entry
TATA-binding protein-associated factor TAF14. DNA binding. RNA polymerase II general transcription initiation factor activity. histone binding. protein binding.
Solver output
TAF14 is a TATA-binding protein-associated factor (TFIID subunit) involved in RNA Pol II general transcription initiation, with additional roles in histone binding and DNA binding. As a general transcription factor component, TAF14 participates broadly in gene expression regulation rather than a specific pro- or anti-longevity pathway. Its histone-binding activity (via its YEATS domain) links it to chromatin state and epigenetic regulation, which is relevant to aging, but the directionality is ambiguous — general transcription machinery components can influence longevity either way depending on which genes they preferentially regulate. There is no strong evidence from its molecular function annotations alone that TAF14 specifically promotes or opposes longevity pathways such as TOR, sirtuins, or stress response in a directional manner.
GenAge curators classify each gene's normal function as promoting (Pro-Longevity) or opposing (Anti-Longevity) longevity, based on synthesis across studies. Loss-of-function effects can vary from this depending on the manipulation.
Advisor judgment
The model predicted "unclear" while the ground truth is "anti_longevity." The model's reasoning is plausible but overly hedged — it correctly identifies TAF14's roles in transcription initiation and histone binding but fails to commit to a direction. TAF14 deletion in yeast extends lifespan (making it anti-longevity), which the model did not predict. The mechanism reasoning about epigenetic alterations is partially relevant (YEATS domain/histone binding) but the model didn't connect this to a specific anti-longevity mechanism. The confidence of 0.38 reflects appropriate uncertainty, but the model still got the answer wrong. The reasoning quality is low-moderate — it identifies relevant biology but fails to leverage it toward a correct prediction.
Raw API responses
solver
{
"reasoning": "TAF14 is a TATA-binding protein-associated factor (TFIID subunit) involved in RNA Pol II general transcription initiation, with additional roles in histone binding and DNA binding. As a general transcription factor component, TAF14 participates broadly in gene expression regulation rather than a specific pro- or anti-longevity pathway. Its histone-binding activity (via its YEATS domain) links it to chromatin state and epigenetic regulation, which is relevant to aging, but the directionality is ambiguous — general transcription machinery components can influence longevity either way depending on which genes they preferentially regulate. There is no strong evidence from its molecular function annotations alone that TAF14 specifically promotes or opposes longevity pathways such as TOR, sirtuins, or stress response in a directional manner.",
"confidence": 0.38,
"key_pathways": [
"RNA Pol II general transcription initiation (TFIID complex)",
"histone/chromatin binding via YEATS domain",
"epigenetic gene expression regulation"
],
"mechanism_class": "epigenetic_alterations",
"longevity_influence": "unclear"
}advisor
{
"notes": "The model predicted \"unclear\" while the ground truth is \"anti_longevity.\" The model's reasoning is plausible but overly hedged — it correctly identifies TAF14's roles in transcription initiation and histone binding but fails to commit to a direction. TAF14 deletion in yeast extends lifespan (making it anti-longevity), which the model did not predict. The mechanism reasoning about epigenetic alterations is partially relevant (YEATS domain/histone binding) but the model didn't connect this to a specific anti-longevity mechanism. The confidence of 0.38 reflects appropriate uncertainty, but the model still got the answer wrong. The reasoning quality is low-moderate — it identifies relevant biology but fails to leverage it toward a correct prediction.",
"failure_mode": "appropriately_uncertain",
"answer_correct": false,
"mechanism_correct": false,
"reasoning_quality": 2,
"ground_truth_questionable": false
}