In the Dec 15, 2016 issue of Cell, three papers from Jonathan Weissman, Aviv Regev, and Ido Amit described this really cool technique called Perturb-seq. This is a technique to do focused CRISPRi screen (of ~100 sgRNAs) followed by Drop-seq. Basically they use CRISPRi to knockdown different genes in different cells, then examine the transcriptome outcome of the different knockdowns through single-cell RNA-seq. Very often after a genome-wide CRISPR screen, you often have many hits, so they can use CRISPRi to validate each hits and explore the mechanism of the KD effect. Although single-cell readout from Drop-Seq could be noisy, they sequence ~500 cells per CRISPRi, so the average from the ~500 should give a much more robust expression readout. Perturb-seq doesn’t have to be a validation step of a CRISPR screen, as I would be totally excited to see the transriptome in MCF7 after knocking down 100 different transcription factors, chromatin regulators, and kinases, as Aviv’s paper beautifully demonstrated. This is super cool!! For a long time we have been searching for a proper application for single-cell RNA-seq in translational cancer research. Perturb-seq really sold me on the practical value of Drop-seq.
Now, before we get too excited, we should examine the raw data from Perturb-seq. Seeing is believing! If the data quality is indeed excellent, then there will be opportunities to develop computational methods for the systematic modeling of gene regulatory networks. Aviv’s paper has some cool informatics modeling, but I am sure there are still good opportunities!!
I got an email the morning after I posted this blog, from my Stanford labmate Serge Saxonov. Turns out he is the CEO of 10X Genomics, the technology platform that enabled the Drop-seq part of Perturb-seq. What a happy surprise it is! I haven’t see Serge since I graduated. He started 23&me right after PhD, and now is the CEO of 10X Genomics, wow!!