i have treataed and untreated sample of bulk_b cells and naive_cd8 cells and i have performed atacseq where i have generated .narrowPeak and .bigwig files. I have a question regarding merge peaks and annotation. Do i have to merge the peak files from both cell types or keep them separate? This is my first time at ATACseq data analysis. I also need to understand the connection with RNAseq

That depends on your analysis. Typically, for differential analysis (and convenience) you would produce a single set of peaks (e.g. by merging, overlapping, etc) and then create a count matrix so you'd have peaks in rows, samples in columns and the raw counts per peak and sample in it. FeatureCounts can do that. That would be the input for downstream analysis. DiffBind has functions to automate all that.