Signal, noise, and sampling: How pool size and replication shape metabolomic inference.

Metabolomics provides direct insight into physiological state, but for small organisms such as Drosophila melanogaster, it typically requires pooling individuals to obtain sufficient material. Pool sizes vary widely across studies with little justification, and the impact pooling and biological replication have on metabolomic characterization and signal detection remains poorly understood. We evaluated the effects of pool size and biological replication on metabolomic profiles and signal detection using two complementary designs in D. melanogaster. First, we tested how pooling (5, 50, or 100 individuals) affects metabolomic structure and reproducibility in inbred and outbred populations. Second, we tested how pool size interacts with replicate number to affect detection of diet-associated metabolite changes under a high-sugar perturbation. Pool size shaped metabolomic profiles: pools of five individuals consistently differed from larger pools, which improved reproducibility in a dataset-dependent manner. In the dietary experiment, smaller pools showed reduced sensitivity, detecting fewer true diet-associated metabolites without increasing false discoveries, and replicate downsampling showed that pool size and replication independently shape signal retention. Detection depended on effect size and variability: metabolites with larger, more stable effects were consistently retained, while smaller more variable effects were rapidly lost under reduced sampling. Beta-binomial modeling confirmed that detection probability reflects a balance between signal strength and measurement variability, with pool size and replicate number independently shaping this relationship. Together, these results show that metabolomic inference depends on the interplay of signal, noise, and sampling design, with pool size and replication jointly shaping the detectability, stability, and interpretation of biological signals.