Comparative Analysis of RNAi Screening Technologies at Genome-Scale Reveals an Inherent Processing Inefficiency of the Plasmid-Based shRNA Hairpin

Abstract

RNAi screening in combination with the genome-sequencing projects would constitute the Holy Grail of modern genetics; enabling discovery and validation towards a better understanding of fundamental biology leading to novel targets to combat disease. Hit discordance at inter-screen level together with the lack of reproducibility is emerging as the technology's main pitfalls. To examine some of the underlining factors leading to such discrepancies, we reasoned that perhaps there is an inherent difference in knockdown efficiency of the various RNAi technologies. For this purpose, we utilized the two most popular ones, chemically synthesized siRNA duplex and plasmid-based shRNA hairpin, in order to perform a head to head comparison. Using a previously developed gain-of-function assay probing modulators of the miRNA biogenesis pathway, we first executed on a siRNA screen against the Silencer Select V4.0 library (AMB) nominating 1,273, followed by an shRNA screen against the TRC1 library (TRC1) nominating 497 gene candidates. We observed a poor overlap of only 29 hits given that there are 15,068 overlapping genes between the two libraries; with DROSHA as the only common hit out of the seven known core miRNA biogenesis genes. Distinct genes interacting with the same biogenesis regulators were observed in both screens, with a dismal cross-network overlap of only 3 genes (DROSHA, TGFBR1, and DIS3). Taken together, our study demonstrates differential knockdown activities between the two technologies, possibly due to the inefficient intracellular processing and potential cell-type specificity determinants in generating intended targeting sequences for the plasmid-based shRNA hairpins; and suggests this observed inefficiency as potential culprit in addressing the lack of reproducibility.

Predicting Cancer-Specific Vulnerability via Data-Driven Detection of Synthetic Lethality

“…As such DAISY cannot be tested by applying it to identify SL interactions in model microorganisms as yeast. Second, DAISY identifies SL interactions based on large scale genomic data and shRNA screens, which are at times noisy and inaccurate (Bhinder et al, 2014). Third, as DAISY is based on the identification of gene inactivation, additional mechanisms of gene inactivation, such as epigenetic and posttranscriptional regulation, should be accounted for in the future.…”

Discovery of a Dicer-Independent, Cell-Type Dependent Alternate Targeting Sequence Generator: Implications in Gene Silencing & Pooled RNAi Screens

“…In a recent report, Ramji and co-workers had made a similar observation pertaining to non-specific outcomes from plasmid based hairpins, contrary to the specific silencing conferred by the siRNA counterpart [30]. So when we compared the silencing sequences from the two screening libraries, we were perplexed to observe a contrasting phenotypic outcome conferred by identical guide sequences [11]. Concentrating on the TRC plasmid-based shRNA hairpins, we decided to theoretically explore the aspect of altered cleavage as a likely factor leading to alternate hits and ultimately such a dissimilar hit list.…”

“…Indeed the technological developments have opened multiple avenues to explore the RNAi screening platform in a broader spectrum. Albeit such progress, the data outputs from RNAi screens have repeatedly failed to reproduce when tested independently [4], [6]–[11]. The upheaval of examples with regards to data discordance has, more than ever, ascertained the need to diligently address the current pitfalls of RNAi data outputs.…”

“…Based on the previous explanations for poor reproducibility, we thought that the best case scenario would be to control for all such environmental variabilities by comparing two screens conducted in the same laboratory, by the same personal, using the same assay, and readouts, and the data to be analyzed by the same methodology; we hoped to observe an excellent overlap amongst the hit lists thus obtained. The only difference in this comparative analysis was the choice of technology, one screen was siRNA based while the other was shRNA based; execution of the siRNA screen had lead to nomination of 1,274 gene candidates and the shRNA had lead to nomination of 497 gene candidates [11], [28]–[29]. The benchmark for these screens was the identification of known genes from the miRNA biogenesis pathway, and it was a matter of concern when DROSHA was the only known modulator identified in the shRNA screen.…”

High-Throughput Silencing Using the CRISPR-Cas9 System: A Review of the Benefits and Challenges

“…48 This phenomenon may have a significant contribution to the low confirmation rates in shRNA screens and the relatively low cross-validation rate between shRNA and siRNA screens. [49][50][51] Given these issues, researchers in the CRISPR/Cas9 field have invested significant efforts in the determination of OTE and their mitigation.…”