I have multiple bam files from cancer datasets and I want to check their quality. What I did was, I converted the bam to fastq using bedtools. I got two fastq read files (paired end) per bam file. Now I checked the quality of the fastq files generated using fastqc tool. This is how I am trying the find a good dataset for my analysis. Is this approach ok?
No. It's not ok.
You are checking the quality of
fastq file but what you want is BAM quality. There is a difference between two. Check Qualimap to assess the quality of aligned BAM.
If it is ChIP-seq data, explore deepTools. It has very nice functions.