Table 1 Experimental and analytical pipelines of the HIPLAB and NIBR datasets
HIPLAB | NIBR | |
|---|---|---|
Number of screens | 3356 | 2725 |
Number of unique compounds | 3250 | 1776 |
Number of HET strains | 1095-essential | 5796-essential+nonessential |
Number of HOM strains | 4810 | 4520 |
Bioassay | IC20 | IC30 |
HIPHOP assay plates/media | 48-well/700ul YPD | 24-well/1600ul YPD |
experiments per plate | 42 drug-treated samples + 6 negative controls (1% DMSO) | 10 drug-treated samples in duplicates + 2 negative controls (no drug) + 1 positive control (Benomyl) + 1 contamination control (no cells) |
HIPHOP assay device | Tecan Genios spectrophotometer | Cytomat Robotic shaking incubator |
starting number of cells | O.D.600 of 0.02 (~ 400 and ~ 200 cells/strain for HIP and HOP respectively) | 100ul and 110ul of a 1.5 O.D.600 /ml culture (~ 600 and ~ 700 cells/strain for HIP and HOP, respectively |
Frequency of Optical Density (O.D.) measurement | 15’ | 60’ |
Collection time | log-phase cells; 20 and 5 generations for HIP and HOP, respectively | saturated cells; ~ 20 and ~ 5 generations for HIP and HOP, respectively |
Final strain intensity value | ‘best tag’: tag with the lowest robust coefficient of variation in the control arrays | average of uptag and downtag intensities |
z-score calculation for straini in screenj | log2 ratioij = log2(median signal from controls/signal from drug-treated sample) | log2 ratioij = log2(average signal from replicates of drug-treated samples/average signal from controls sample) |
z scores = FDij = MADL = (log2 ratioij - median of log2 ratioj) / MAD log2 ratio screenj | Sensitivity scores = FDij = MADL = (log2 ratioij - median of log2 ratioj) / MAD log2 ratio screenj | |
Adjusted MADL based on the variability between replicates: aMADL = min(0.05/p,1)*MADL | ||
z scores = aMADL/standard deviation of aMADL values of strain i over n screens | ||
Significant chemical-genetic interactions | standard normal distribution P ≤ 0.001 | z-score < -5 |
Clustering method | Ward hierarchical clustering with dynamic branch cutting | Average-linkage two-way hierarchical clustering |