Dataset features


Dataset type: Other
Number of samples: 25
Release date: Feb 10 2012
Last update date: Jan 18 2013
Access: Public
Chemicals: Adenosine Monophosphate
Dataset link A genome-wide genetic interaction screen reveals metabolic roles of human histone deacetylase 1

Experimental Protocol

To study the functional specificity of individual HDAC, we developed a genome-wide genetic interaction profiling technology in cultured human cells by RNAi using pooled TRC (The RNAi Consortium) 75k human shRNA library and complexity deconvolution by high-density microarray with a half-hairpin barcode design. The HCT116 colon cancer cell line was chosen as the screen platform because of its quasinormal diploid karyotype. The performance of the customized microarray we designed was first evaluated by the receiver operating characteristic (ROC) curve showing high sensitivity (89%), specificity (94%) and area under curve (0.95), as well as a control hybridization without DNA sample showing low background signal. Both results suggest that the fluorescent signals were generated by specific hybridization reactions. The high correlation of signal between Cy5 and Cy3 channel as well as dye-swap experiments indicated that biases between the two labeling fluorophores during competitive hybridization were negligible. Moreover, correlation between technical and biological replicates confirmed the high reproducibility of the methodology. In the screen, we used stable polyclonal cells expressing shRNA constructs targeting firefly luciferase (shLuciferase) as control. We checked the knockdown efficiency in protein or RNA level of individual shRNA constructs for 12 human HDAC genes (HDAC1~4, HDAC6~9, SIRT1~3 and SIRT5) by immunoblotting or quantitative PCR, and generated stable polyclonal query cell lines expressing two shRNA constructs with the highest knockdown efficiency for each HDAC gene. After transduction by TRC shRNA lentiviral pools, benchmark samples were harvested prior to puromycin selection, and the remaining cells were propagated under selection and harvested again after 18 population doublings as the end samples. Half-hairpin barcode library of the benchmark and end samples was recovered from genomic DNA respectively by PCR with Cy5 and Cy3-labeled primers, gel purified, and hybridized to the microarray. For each shRNA construct, the Z-score of the log2(Cy5/Cy3) was computed, and Z-score difference was calculated by subtracting the Z-score of the control sample from that of the query sample. Z-score differences larger than 1.5 and less than -1.5 were used as arbitrary thresholds to define candidate negative and positive genetic interactions, respectively.










Yu-yi Lin
Yu-yi Lin

Dataset Statistics


Citations per year

Dataset publication