From: A Multi-feature Reproducibility Assessment of Mass Spectral Data in Clinical Proteomic Studies
Distribution of bias with different parameter | Number of samples–replicates grouped in the same cluster | Discordance index (% of samples–replicates that failed to group in the same cluster) | Median % change across all features |
---|---|---|---|
Normal (n = 30) | |||
µ = 0 σ = 4 | 27 | 0.10 | 0.4% [−24%, 21%] |
µ = 2 σ = 2 | 27 | 0.10 | 67% [16%,288%] |
µ = 2 σ = 4 | 23 | 0.23 | 78.2% [17.1%, 313.2%] |
Exponential (n = 30) | |||
\( \lambda = {\text{1}}\left( {\mu = \sigma = {\text{1}}} \right) \) | 30 | 0.0 | 43% [8.6%, 137%] |
\( \lambda = 2\left( {\mu = \sigma = 0.5} \right) \) | 30 | 0.0 | 21% [4.3%, 68%] |
Bimodal (n = 30) | |||
\( \mu = 1,\sigma = 2\left( {m/z \leqslant 1,000} \right)\mu = 2,\sigma = 4\left( {m/z > 1,000} \right) \) | 26 | 0.13 | 65%[8.8%,191%] |
\( \mu = 2,\sigma = 4\left( {m/z \leqslant 1,000} \right)\mu = 4,\sigma = 8\left( {m/z > 1,000} \right) \) | 10 | 0.67 | 130% [18%, 381%] |