how to caculate rpkm for bacteria RNA-SEQ data?

[joyjane88]

hello, everyone! I am a new member here.
Recently, I am working with the bacterial RNA-SEQ data. My data was in strand-specific PE reads fq file. I have mapped them onto the reference genome sequence and have got the sorted and indexed bam file. Nextly,
I want to make a differential gene expression analysis. So, I want to know what tools is recommended by the hands-on. DEseq, EdgeR, Cuffdiff or any other one, which is better?
Any opinions or Suggestions will be appreciated. Thank you!

[dpryan]

I've generally been happy with DESeq2. As to which is better, there's really no single answer to that. They all use slightly different methods, so it's not unheard of for one package to model experiment A (but not B) better than the others.

[joyjane88]

I've generally been happy with DESeq2. As to which is better, there's really no single answer to that. They all use slightly different methods, so it's not unheard of for one package to model experiment A (but not B) better than the others.

Do you mean that all these tools are suitable for my data and will produce very similar results? I have seen a paper shown that DEseq identify more DEGs than EdgeR when the authors dealing their data with these two pakages however they select the EdgeR result for subsequent study, can I come to a conclusion that DEseq is more senstive?

[dpryan]

Do you mean that all these tools are suitable for my data and will produce very similar results? I have seen a paper shown that DEseq identify more DEGs than EdgeR when the authors dealing their data with these two pakages however they select the EdgeR result for subsequent study, can I come to a conclusion that DEseq is more senstive?

Yes, they'll all work for your data and probably produce pretty similar results. I would expect that DESeq2 will produce a little better results than the others (including DESeq1). I would say that DESeq2 is probably more sensitive, simply due to the shrinkage methods. Aside from that, they're all pretty similar.

[joyjane88]

Yes, they'll all work for your data and probably produce pretty similar results. I would expect that DESeq2 will produce a little better results than the others (including DESeq1). I would say that DESeq2 is probably more sensitive, simply due to the shrinkage methods. Aside from that, they're all pretty similar.

Thanks for your suggestions and explaination! I will undertake my work with DEseq2.