Hi, I tried to use the gene_summary.txt file by fo...
# crisprseq
a
Hi, I tried to use the gene_summary.txt file by following https://bioconductor.org/packages/release/bioc/vignettes/MAGeCKFlute/inst/doc/MAGeCKFlute.html tutorial and when I do log10 of the FDR of the treated, I get no results. My goal is to get the enriched and depleted genes. What would be the best tool for this kind of analysis?
Copy code
But when I use: 

gdata_cc$Control = rowMeans(gdata_cc[,ctrlname, drop = FALSE])
gdata_cc$Treatment = rowMeans(gdata_cc[,treatname, drop = FALSE])

p1 = ScatterView(gdata_cc, "Control", "Treatment", label = "Gene", 
                 auto_cut_diag = TRUE, display_cut = TRUE, 
                 groups = c("top", "bottom")
                )
print(p1)
l
Hi Aastha, did you have a look at the quality of your screens beforehand?
a
Hi @Laurence Kuhlburger, just looked at the multi-QC plots and it looks the sample quality isn't great. If I understand correctly, beta score from the mageck mle is the logFC and can be used to plot the p val and logfc plot to show the enriched and depleted peptides, correct? I have two samples and each have controls. T1_1 vs control1, T1_2 vs control1 and T2_1 vs control2, T2_2 vs control2. design matrix:
Copy code
Samples	        control_1 control_2         T1_1     T1_2     T2_1      T2_2		
control_1	      1            0	          0	      0	        0	      0
control_2	      1            1	          0		  0		    0	      0
T1_1	          1	           0	          1       1	        0	      0
T1_2	          1            0	          1		  1		    0	      0
T2_1	          1            0	          0		  0		    1	      1
T2_2	          1            0	          0		  0		    1         1
Can you help me figure out if there can be any way i can verify everything so that I know i am not doing anything wrong with the pipeline? Thank you
here are some plots
then i plotted as a volcano plot and it doesn't look correct.
Copy code
p <- ggplot(data, aes(x = T1_1.beta, y = T1_1_log10_pvalue)) +
  geom_point(aes(color = significant)) +
  scale_color_manual(values = c("Up" = "red", "Down" = "blue", "Not Sig" = "grey")) +
  geom_vline(xintercept = c(-1, 1), linetype = "dashed") +
  geom_hline(yintercept = -log10(0.05), linetype = "dashed") +
  theme_minimal() +
  labs(title = "Volcano Plot ",
       x = "Log2 Fold Change (which is beta score)",
       y = "-Log10 P-value") +
  theme(legend.position = "right")
l
hey what does the gini index in your multiqc say? You seem to indeed have QC issues 😕