Convergent relaxation of molecular constraint in herbivores reveals the changing role of liver and kidney functions across mammalian diets

Table 3.

Phenotype-based gene sets from the Mammalian Phenotype Ontology that are significantly enriched for genes associated with carnivory score

Phenotype Enrichment statistic Parametric P-value Parametric Q-value Permulation P-value Permulation Q-value
MP:0002223 lymphoid hypoplasia −0.304 8.60 × 10−4 0.076 <7.00 × 10−5 a <0.057a
MP:0003949 abnormal circulating lipid level −0.235 3.30 × 10−5 0.011 7.00 × 10−5 0.057
MP:0001547 abnormal lipid level −0.170 3.62 × 10−9 1.20 × 10−5 8.00 × 10−5 0.057
MP:0005441 increased urine calcium level −0.311 7.74 × 10−8 8.53 × 10−5 8.00 × 10−5 0.057
MP:0002988 decreased urine osmolality −0.181 1.01 × 10−5 0.006 8.00 × 10−5 0.057
  • Permulation P-values represent the proportion of 100,000 permulations that produced a stronger enrichment statistic than the observed value for each gene set. Multiple hypothesis testing corrections were performed by generating Q-values using Storey's correction method (Storey et al. 2020) (FDR = 0.1). A negative enrichment statistic signifies the following pattern: the greater the decrease in carnivory, the higher the rate of evolution of the gene.

  • aAfter generating 100,000 null statistics, none produced a stronger correlation with diet than the observed values for MP:0002223. However, if the P-values are adjusted to the smallest observed nonzero P-value (7.00 × 10−5), they would produce a significant empirical Q-value (FDR = 0.1).

This Article

  1. Genome Res. 34: 2176-2189

Preprint Server