Consensus Scores

An indicator of data robustness and reproducibility. Do not rely on a single p-value.

💡Why Consensus Matters?
Q-omics evaluates the robustness of data associations by testing them across diverse analytical conditions. A high consensus score reduces false positives caused by specific sample splits or cut-offs.
Sampling Consensus (Technical)

How consistently is the association observed when changing analysis parameters (e.g., sample splitting, gender, stage)?

Lineage Consensus (Biological)

Is this association specific to a single cancer type (Unique) or conserved across multiple tumor lineages (Pan-cancer)?

Example 1: Gene Expression & Patient Survival

TNMD Table Example

Interpretation of the Score (31 / 144)

  • The Scenario: TNMD expression shows a significant association with patient survival (OS) in ACC females (Stage II-IV).
  • Sampling Consensus (31): This association was reproducible in 31 out of 144 different sampling subgroups.
  • Denominator (144): Derived from all possible combinations of: Data split (Median/Quartile), Gender, Stage, Measures (OS/DFS) & Periods.

Lineage Consensus (10)

  • In 10 tumor lineages out of 34, TNMD expression is significantly (p<0.05) associated with patient survival in at least one sampling group.

Example 2: Drug Response & Gene Expression

Dasatinib Table Example

Interpretation of the Score (2 / 4)

  • The Scenario: Dasatinib response is correlated with GSTP1 expression in Breast cancer cell lines.
  • Sampling Consensus (2): Significant correlation was found in 2 out of 4 analytical conditions.
  • Denominator (4): Combinations of data splitting methods for both Drug response and Gene expression (Median vs. Quartile).

Lineage Consensus (3)

  • This drug-gene association is observed in 3 out of 20 different cancer lineages (Pan-cancer potential).