Sources of Error in Mammalian Genetic Screens

Abstract

Genetic screens are invaluable tools for dissection of biological phenomena. Optimization of such screens to enhance discovery of candidate genes and minimize false positives is thus a critical aim. Here, we report several sources of error common to pooled genetic screening techniques used in mammalian cell culture systems, and demonstrate methods to eliminate these errors. We find that reverse transcriptase-mediated recombination during retroviral replication can lead to uncoupling of molecular tags, such as DNA barcodes (BCs), from their associated library elements, leading to chimeric proviral genomes in which BCs are paired to incorrect ORFs, shRNAs, etc. This effect depends on the length of homologous sequence between unique elements, and can be minimized with careful vector design. Furthermore, we report that residual plasmid DNA from viral packaging procedures can contaminate transduced cells. These plasmids serve as additional copies of the PCR template during library amplification, resulting in substantial inaccuracies in measurement of initial reference populations for screen normalization. The overabundance of template in some samples causes an imbalance between PCR cycles of contaminated and uncontaminated samples, which results in a systematic artifactual depletion of GC-rich library elements. Elimination of contaminating plasmid DNA using the bacterial endonuclease Benzonase can restore faithful measurements of template abundance and minimize GC bias.

A Multiplexed Single-Cell CRISPR Screening Platform Enables Systematic Dissection of the Unfolded Protein Response

“…In engineering our vector, we first incorporated three tandem sgRNA expression cassettes (composed of an RNA polymerase III promoter, sgRNA protospacer, and sgRNA constant region) into our Perturb-seq vector (Figure 2A). To minimize intramolecular recombination at repetitive nucleotide sequences during lentiviral transduction (Sack et al, 2016; Smyth et al, 2012), we used three different promoters in this initial three-guide vector (Methods). Test vectors expressing sgGFP from one of the promoters (and negative control sgRNAs from the others) partitioned GFP+ K562 cells with dCas9-KRAB into two subpopulations with either strong GFP depletion (>90%) or no detectable depletion (Figure S2A, S2B, Methods).…”

“…Cost per cell will decline as technologies mature, and sequencing costs can be mitigated through amplification of select targets (like our guide-mapping amplicons) or depletion of uninteresting high abundance genes (Gu et al, 2016). A more subtle point is that intermolecular provirus recombination during transduction can scramble barcode identities in pooled lentivirus preparations (Sack et al, 2016). We took careful steps to avoid this problem and expect that straightforward protocol alterations will circumvent this issue.…”

“…Viruses were individually packaged (using sequence-verified lentiviral Perturb-seq vectors or final three-guide Perturb-seq vectors) and harvested in preparation for Perturb-seq screening. Individual packaging of the lentivirus and pooling at the step of virus or cells was done to avoid intermolecular recombination of proviral genomes and to ensure maintenance of paired barcode-sgRNA coupling (Sack et al, 2016). For the “pilot experiment” (schematic in Figure S1A, data represented in Figures 1C–E, S1B) cBA010 cells were individually spinfected with virus (at 33°C for 2 hours at 1000× g ) in media supplemented with 8 μg/mL polybrene; 5 hours post spinfection, virus was removed by centrifugation and cells were resuspended in fresh media.…”

Protein Barcodes Enable High-Dimensional Single-Cell CRISPR Screens

“…To test this possibility, we generated 96 CRISPR gRNAs targeting 54 different genes (1–3gRNA/gene), and paired each gRNA with a different Pro-Code. As it has recently been reported that packaging vector pools together can lead to varying degrees of barcode swapping (Hill et al, 2018; Sack et al, 2016), we made each vector individually, and subsequently pooled them in equimolar ratio, as this eliminates the possibility of swapping (Adamson et al, 2016). THP1 human monocytes were engineered to express Cas9 (THP1-Cas9), and transduced with all 96 Pro-Code/CRISPR vectors together as a pool.…”

Mapping Cellular Reprogramming via Pooled Overexpression Screens with Paired Fitness and Single-Cell RNA-Sequencing Readout

“…Since barcode shuffling has been identified as a key factor limiting the power and sensitivity of single-cell based screening (Adamson et al, 2016, 2018; Sack et al, 2016; Hill et al, 2018; Xie et al, 2018), we also explored an alternate version of the overexpression vector (TF-NoHygro, Figure S1c) to minimize template-switching events during lentiviral packaging. The rate at which the association between the ORF and barcode is lost due to template switching is proportional to the length of the constant region between the ORF and barcode, which in this case is the selection marker.…”