Verification

We estimated the false positive rate by performing chromatin immunoprecipitation followed by quantitative gene-specific PCR analysis on sets of genes predicted to be bound by TAL1 in Jurkat cells. We designed primers suitable for quantitative PCR in the promoter region of the genes identified as TAL1 targets by ChIP on chip. Then we amplified those regions and tested for their enrichment in the immunoprecipitates (IPs) vs. whole cell extracts by quantitative real-time PCR using -actin as our control gene. We considered that a promoter region was validated as target for TAL1 when it is enriched at least two fold in the IP compared to the whole cell extract. The complete list of validation results is included in Table S2 in the Supplemental Materials.

We use an error model to provide confidence in the significance of the enrichments we observe. To enable cogent discussion of the results, we selected a threshold (p<0.001) that was intended to minimize false positives, though at the expense of increasing false negatives. Additional binding interactions may be identified by applying less stringent thresholds. Our quantitative gene specific PCR on TAL1 chromatin immunoprecipitates show a validation rate for TAL1 direct targets identified by ChIP on chip of approximately 80%, which is comparable to the rates observed for other transcription factor and tissues using the same Hu13K array (see Odom et al, 2004). Non verified targets may correspond to false positive hits in the ChIP on chip analysis but also to promoters bound by TAL1 in regions difficult to amplify by PCR where quantitative PCR assays fail.

Chromatin immunoprecipitations for TAL1 were performed in non-synchronized Jurkat cell cultures. Thus, the targets bound by TAL1 are independently of the cell-cycle stage. Jurkat cells are an immortalized tumor line with a very high expression of TAL1 both at the RNA and at the protein level, similar to that detected in human primary T-ALL samples; therefore, we expect that most of the TAL1 targets identified here reflect the activity of TAL1 in primary leukemias.