I am currently using Deseq2 for RNA seq analysis. I have two groups with three biological replicates w/ a control and KO mouse line. Here is what the PCA plot currently looks like I am removed low count from the data but didn't do any normalization steps. Any advice would be great!

Code should be placed in three backticks as shown below

CT2_dds <- DESeqDataSetFromMatrix(countData = countData_2, colData = colData_2, design = ~ condition) 
keep_2 <- rowSums(counts(CT2_dds)) >= 10 
CT2_dds  <- CT2_dds[keep_2,]

CT2_dds <- DESeq(CT2_dds)

CT2_res <- results(CT2_dds, alpha = 0.05)


# include your problematic code here with any corresponding output 
# please also include the results of running the following in an R session 

sessionInfo( )

You should check your data for batch effects before running deseq2. The wild type sample way out on its own is clearly an issue

1) Known batch effects - these are known technical variables in your metadata (e.g. age, sex, data of sample extraction, processing date etc.). They should be added to the design parameter of deseq2 (e.g. if some samples are male and others are female, then the design should include sex)

2) Hidden batch effects - these are unknown variables that can be assessed with an RLE plot, and then modelled with SVA/RUV for addition to the design parameter of deseq2

3) I also annotate my counts data with biotype so I can plot a biotype v tpm plot for each sample. It is a handy way of seeing if I have contamination with a biotype that shouldn't be there e.g. rRNA, mtRNA

In my experience, when the first PC is 91% of the difference, its because you are looking at two different tissues. Check to see what genes are driving PC1, see if they make any sense for your experiment. You are going to have to drop that sample, it would be a lot nicer if you could explain why (like "this sample is contaminated with muscle")