Gene expression profiling of Drosophila S2R+ cells following RNAi-mediated knockdown of transcription factors

In order to study the effect of transcription factor knockdown, we selectively depleted mRNA products from 483 different genes that are known or predicted to encode transcription factors. We treated Drosophila S2R+ tissue culture cells with double strand RNAs designed to be specific for these loci. Following RNAi treatment, we isolated poly A+ RNA from the cells, and performed stranded high-throughput RNA-Seq analyses to determine knockdown efficiency and propagating transcriptional consequences. Overall design: We started with a curated list of transcription factors and related proteins compiled for production of a reagent library at the DRSC (http://www.flyrnai.org/supplement/TranscriptionFactorGenes.xls). We ranked genes on this list as follows - Rank 1: experimental evidence of TF activity and DNA binding, Rank 2: experimental evidence of TF activity and predicted DNA binding, Rank 3: predicted/unknown TF activity and experimental evidence of DNA binding, Rank 4: predicted/unknown TF activity and/or predicted DNA binding, Rank 5: experimental evidence of TF cofactor, Rank 6: predicted TF cofactor, Rank 7: chromatin regulation, Rank 8: little or indirect annotation, Rank 9: no evidence for any TF activity. We then excluded genes with a rank of 8 or 9. We additionally restricted the list to the subset of 483 genes with expression levels higher than signal from intergenic regions in an Drosophila S2R+ RNA-Seq study (Lee et al., 2014. PMID: 25262759). We generated a panel of double strand RNA (dsRNA) for knockdown reagents (Ramadan et al., 2007. PMID: 17853882). For 387 loci we used two different dsRNA reagents, which target different regions of the mRNAs. Only one dsRNA reagent was used for the remaining 96 loci. As a control for general effects of dsRNA treament, we used dsRNA against the E. coli LacZ gene as a sham. We performed RNAi treatments and library preparations in 96 well format, and we placed one LacZ control in each plate. On each plate, there were also wells that we used to estimate dsRNA uptake after depletion of Death-associated inhibitor of apoptosis 1 (Diap-1, or thread. Flybase ID: FBgn0260635) and ensured that cells underwent cell death. RNA-Seq libraries from wells with such high proportions of dead cells were not informative and are not part of this submission. We performed all dsRNA treatments in biological duplicate. We performed RNA-sequencing on Illumina platform after polyA+ selection of mRNA. We prepared libraries in strand-specific manner, using the Uracil-DNA glycosylase method (Parkhomchuk et al., 2009. PMID: 19620212). RNA-Seq libraries were sequenced to generate single-end, 50 bp reads.