Journal Club

[alsi...@gmail.com]

Dear Atray, 

I'm very interested in your paper. Since I will have to expose your results to my colleagues and to younger pupils - I am a university master degree student - as an exercise, I wondered if a more detailed and clarifying version of the paper is available, so as to have a clearer view of the workflow and the graphs obtained, since the statistics you use are rather complex and that we have not yet attended a course like that!

Best regards,

Alessio

PS. If you do not have such a thing, do you mind if I write you a list of questions, albeit trivial?

- What do you mean for autonomous phenotype?

- Why did you choose those specific cell lines to demonstrate Perturb-seq?

- So, basically, at first you have generated the plasmid and created the libraries matching each sgRNA with the GBCs (I don't get why you have done the NGS to pair sgRNA with GBC, I mean didn't you know both their sequences? Did you perform it as a control?); then you have infected different cells and in the case of BMDC you have stimulated them with LPS. then, performing the scRNA-seq, you have tagged each cell's mRNA with the CBC (so, all the mRNA of one cell have a specific CBC) and a unique UMI (each one specific for each mRNA moleucle) and finally thanks to the CBC you can distinguish each cell's transcriptome in which you have the tagged mRNAs thanks to the GBG, detected thanks to the final NGS, so that in this way you can understand where the perturbation occurred. Is it correct?

- Then I don't get the correlation between the probability of GBC detection an MOI; why the predicted fit was undistinguishable? 

- In the MIMOSCA, what do you mean for the regulatory effect of each guide on each gene? Is it the KO?

- What do you mean for covariates? And I don't get pretty well the filtering process. Can you explain with more details the Figure 2? What are the residuals of the model?

- What do you mean for "Perturb-seq recovers the correct genes" if you are performing a KO?

- Did you cultured with GM-CSF to obtain more precursors?

- Why did you use so many guides to promote the KO for just 24 TFs? It's not 1:1?

- In the figures S3E and S3D you are comparing to the nontargeting condition (which i think is the control)?

- Did you stimulate with LPS to induce a certain DCs maturation? 

- Figure 3C

- What are the GO terms?

- Figure 5D, E, F: how did you obtained the 27 categories? By combining the 3 possible states? And did you analysed pairs of genes since they are those significant among the paired sgRNAs that you have detected? And the last point, what is the meaning of the F picture?

- What are the 2 modules in the 6A picture? 

- How did you defined the nine cell states? The 6C is a gene ontology, too?

- What are the PC scores? And how can you say that the TF-specific effects are reproducible?