How should I count the percentage of messenger RNA and ribosomal RNA in the samples

[illuminaGA]

Dear All

We have some E coli. total RNA-seq data and my PI would like to count the percentage of messenger RNA and ribosomal RNA in the data. My idea is mapping the data to the transcriptome and rRNA data separately and then count the reads#.

Do you think this is a doable plan? Actually, I have no idea where to find the transcriptome and rRNA reference for the E coli.

Is there is an alternate way to do this analysis? Any suggestion will be appreciated.

Thanks.

[GenoMax]

You can find the rRNA sequences for E. coli here. You can align rest of the data to the genome and count the reads using a GTF file. I don't recollect if E. coli has overlapping reading frames but otherwise it should be straight forward to do the counts.

[illuminaGA]

You can find the rRNA sequences for E. coli here. You can align rest of the data to the genome and count the reads using a GTF file. I don't recollect if E. coli has overlapping reading frames but otherwise it should be straight forward to do the counts.

Thank you so much.

Btw, how / where should I download the rRNA sequences? I explore the database for hours but still cannot figure out where to download the sequence.

Thanks again.

Al

[GenoMax]

Use the links I included above. Then click on "nucleotide sequence" in the operations panel to the right.

[illuminaGA]

Use the links I included above. Then click on "nucleotide sequence" in the operations panel to the right.

Great, Thanks,

So what should I do is copy those sequences into a text file and build as a reference and map the sample sequencing data to this reference, right?

One more question, can I just use the mappable reads number as the rRNA reads number?

Thanks a lot again.

AL

[illuminaGA]

Use the links I included above. Then click on "nucleotide sequence" in the operations panel to the right.

Dear GenoMax

I just finished the mapping to the rRNA reference and got 70% mapping rate. Is this a normal range for the mRNA seq? Can I say 70% reads are rRNA? how to understand the result? Could you please give me some pointers? Thank you so much.

AL

[GenoMax]

You say mRNA seq but had you done any ribosomal RNA depletion (e.g. https://www.illumina.com/products/by...-bacteria.html) on your samples? If not, it is not surprising to see a large fraction of your sample to be rRNA. Unless you are working with rRNA that part of the sequence data is wasted (reason to do ribo-depletion).

[illuminaGA]

You say mRNA seq but had you done any ribosomal RNA depletion (e.g. https://www.illumina.com/products/by...-bacteria.html) on your samples? If not, it is not surprising to see a large fraction of your sample to be rRNA. Unless you are working with rRNA that part of the sequence data is wasted (reason to do ribo-depletion).

Oh no. I will confirm with the lab to see what kit they use. how should I remove those reads from the raw reads? I saw someone said it's not necessary to do that.

Appreciate your help again.

AL

[GenoMax]

You could extract the unmapped reads from the alignment you did (if you did include them in your alignment file) or redo the alignment and collect the unmapped reads in a separate file.

You could also ignore these reads when you do read counts. You would want to compare samples and make sure rRNA contamination levels are more or less the same across your pool of samples. You don't want one sample to have 70% rRNA and other 5% (if total number of reads are more or less similar).

[illuminaGA]

You could extract the unmapped reads from the alignment you did (if you did include them in your alignment file) or redo the alignment and collect the unmapped reads in a separate file.

You could also ignore these reads when you do read counts. You would want to compare samples and make sure rRNA contamination levels are more or less the same across your pool of samples. You don't want one sample to have 70% rRNA and other 5% (if total number of reads are more or less similar).

Thanks a lot. I mapped the reads by Tophat, one of the output files is unmapped.bam. Can I just convert this file to a fastq file and map again?

Another sample mapping is still running, I will compare the % of rRNA.

Thanks again for your help

AL

[GenoMax]

You should stop using TopHat for new projects. Use BBMap, STAR, HISAT2 etc.

[illuminaGA]

You should stop using TopHat for new projects. Use BBMap, STAR, HISAT2 etc.

Ok, Thanks a lot.

Is there an alternative software for cufflink? It seems very slow.