Literature reference bars

Ref

baseline-literature-reference-bars

Reference bars are literature ballparks under family-holdout: Traitar (Weimann 2016) for categorical traits, Sauer & Wang (2019) for OGT, Weissman 2021 for growth-rate-adjacent codon signals, dbCAN / CAZy for carbohydrate pathways, and marker-gene analyses for N/S/Fe cycles.

Composite

—

task headline

Wall clock

—

Compute

—

Agent cost

—

Started

—

Results by dataset

Raw metric and 95% CI where reported, next to the dataset's chance level, reference and SOTA.

Dataset	Metric	Value	95% CI	Chance	Ref.	SOTA	n · coverage
oxygen tolerance Growth conditions	Macro F1	0.700	—	0.250	0.700	—	— · —
temperature range Growth conditions	Macro F1	0.600	—	0.250	0.600	—	— · —
optimum temperature c Growth conditions	Spearman ρ	0.700	—	0.000	0.700	—	— · —
optimum ph Growth conditions	Spearman ρ	0.350	—	0.000	0.350	—	— · —
gram positive Morphology & dispersal	AUROC	0.950	—	0.500	0.950	—	— · —
sporulation Morphology & dispersal	AUROC	0.900	—	0.500	0.900	—	— · —
motility Morphology & dispersal	AUROC	0.850	—	0.500	0.850	—	— · —
cell shape Morphology & dispersal	Macro F1	0.550	—	0.143	0.550	—	— · —
nitrogen fixation Metabolic pathways	AUPRC	0.900	—	0.061	0.900	—	— · —
nitrate reduction Metabolic pathways	AUPRC	0.750	—	0.597	0.750	—	— · —
denitrification Metabolic pathways	AUPRC	0.750	—	0.034	0.750	—	— · —
sulfate reduction Metabolic pathways	AUPRC	0.800	—	0.011	0.800	—	— · —
fermentation Metabolic pathways	AUPRC	0.650	—	0.152	0.650	—	— · —
cellulose degradation Metabolic pathways	AUPRC	0.750	—	0.065	0.750	—	— · —
xylan degradation Metabolic pathways	AUPRC	0.700	—	0.042	0.700	—	— · —
chitin degradation Metabolic pathways	AUPRC	0.700	—	0.015	0.700	—	— · —
fumarate reduction Metabolic pathways	AUPRC	0.550	—	0.038	0.550	—	— · —
iron reduction Metabolic pathways	AUPRC	0.600	—	0.053	0.600	—	— · —

Integrity

Reference result.

Clean

provenance
Reference bars are literature ballparks under family-holdout: Traitar (Weimann 2016) for categorical traits, Sauer & Wang (2019) for OGT, Weissman 2021 for growth-rate-adjacent codon signals, dbCAN / CAZy for carbohydrate pathways, and marker-gene analyses for N/S/Fe cycles.
pass

Run details

Kind: baseline
Suite: bacdive_multitrait v1
Started: —
Finished: —
Artifacts: —