GeCKO Analysis

[Elizabeth S]

Hi everyone,

Has anyone used one of the free analysis tools to analyze their GeCKO screen?  I've been looking into MAGeK, edgeR, and CRISPResso, and would appreciate any feedback about the ease of use and quality of these tools. 

Thanks for your help,

Elizabeth 

[Mark]

Hi Elizabeth,

So far I've tried RIGER and MAGECK. I like the latter for a couple of reasons:

1. You can either start with your FASTQ files, raw read count, or normalized read count tables for analysis;

2. You can enter your replicate experiments for your input and it automatically takes them into account for the gene depletion/enrichment analyses.

3. Their raw read count normalization method in my opinion is better than the manual normalization method (% of total # of reads). I find more of my negative control guides in my hits when I normalize manually. 

4. It has a visualization tool that generates publication-quality figures.

5. They're very quick and patient to respond to questions RE: troubleshooting/editing the program. 

6. Their YouTube tutorials are very helpful even if you have the most basic command line experience.

The Moffat lab has release a new algorithm as well. Works slightly better than MAGECK when it comes to lower ranking hits. Otherwise, they're very much similar with regards to your best hits. 

http://www.sciencedirect.com/science/article/pii/S0092867415014956

http://biorxiv.org/content/early/2015/11/27/033068

Haven't really tried running BAGEL so couldn't comment on the ease of using it.

Hope that helps!

[Elizabeth S]

Thanks, Mark!  I think we are going to try MAGECK.  

Elizabeth 

[Katherine Shortt]

Hi all, 

I am currently working on this part of the analysis in my GeCKOv2 screen and am comparing MaGeCK to bowtie2+something to generate read counts+RIGER.  Which normalization method in MaGeCK do you prefer?  If you generate read counts and normalize them outside of MaGeCK what software or method are you using?

Thanks for the advice!