Question: Mycoplasma Cell Culture Contamination Affecting High-Throughput Data [Chip-Seq & Rna-Seq]
gravatar for Sukhdeep Singh
6.7 years ago by
Sukhdeep Singh9.6k
Sukhdeep Singh9.6k wrote:

Hi guys, we recently have discovered Mycoplasma contamination in the cell culture, and an effort is being done to remove it. I was wondering, if I can check the previous ChIP-seq & RNA-Seq samples which might be effected by this. I have want to add Mycoplasma penetrans and Mycoplasma genitalium to the list, if we have lot of reads (not aware of how much should be the threshold) mapping to the genome, it means there is a lot of background from the Mycoplasma, else its fine.

So, do you guys think, I should add some other Mycoplasma species as well, which might commonly effect the cell cultures, as I don't know how similar the Mycoplasma sub-species genome. Also, should I care about the ChIP-Seq contamination with Mycoplasma (may the mycoplasma reads dont map to the mouse genome). The antibody used for the protein pulldown is the anti-GFP.

Are there any other measures or tools to test that, like providing the Genome file as an parameter and the reads mapping to it, should be discarded and the rest should be proceeded with mapping to the genome of organism of interest.



chip-seq rna-seq • 3.8k views
ADD COMMENTlink modified 6.6 years ago by Michael Dondrup45k • written 6.7 years ago by Sukhdeep Singh9.6k
gravatar for Michael Dondrup
6.7 years ago by
Bergen, Norway
Michael Dondrup45k wrote:

I believe any sequencing approach should, as a standard of the QC step, involve scanning for contaminants. I am not sure to which level the ChIP-seq results can be affected but the RNA-seq results most likely will be. In any case the bacterial contamination will disturb your experimental parameters and therefore the results from contaminated cell-assays should be generally discarded (my oppinion). I know from biologists that it is hard to keep their stuff sterile but it is definitely possible.

ADD COMMENTlink written 6.7 years ago by Michael Dondrup45k

But what genome to choose to scan contamination, are there any common ones and we are not sure, the samples the effected or not. Any specific tool or way to test that except mapping??

ADD REPLYlink written 6.7 years ago by Sukhdeep Singh9.6k

The biologist I asked said it is often mycoplasma and also yeast. Maybe scan all reads that don't align to human against bacterial and yeast, or blast against NT even.

ADD REPLYlink written 6.7 years ago by Michael Dondrup45k

So, thats the thing, if they do not align, then we don't care, if the read density is enough (>20 Mil). just for ref, I tried one sample, only 3 reads mapped out of 43 Million to the Mycoplasma penetrans genome. Thanks for the comments

ADD REPLYlink written 6.7 years ago by Sukhdeep Singh9.6k

I think there are 2 issues here. 1. Reads aligning to the genus mycoplasma. As can be seen here about 20 species of Mycoplasma could be associated with contaminated cell cultures. It might not be a problem to scan your reads against these genomes. And it is upto you if are willing to tolerate what you see - like you said, 3 reads out of 43 Mi mapped to M.penetrans genome. 2. Effect of mycoplasma contamination on the expression profile you are investigating. How sure can one be that the profile you see in your sample is the real effect and not because of Mycoplasma contamination (maybe I am making a mountain out of mole .. but just sharing thoughts!!). In fact this approach is currently being used to study the expression profiles of mixed populations. for eg. Mammalian cells infected with bacteria, the infected cells seperated by FACS or other methods and the mixed RNA isolated from the mixture of host and bacteria are subjected to sequencing and processed by mapping on to the respective genomes. Here the setup is meant to be that way while in your case it is undesired that you have mixed RNAs in your sample. As a solution, how about looking for some markers / house keeping genes to make sure that the mycoplasma has not done anything drastic to your cells with respect to altering expression profile from your dataset?

ADD REPLYlink written 6.7 years ago by abi280

Good points abi, ChIP-Seq, i think no problem, in peak calling also, if the contamination is not that much, they will be ruled out. Also, a new mock will be better to use, if its contaminated as well. For RNA-seq, its little more complicated, as you pointed out. I will look into that, genes which are regulated by Mycoplasma, may or may not be regulated in mouse. Thanks

ADD REPLYlink written 6.7 years ago by Sukhdeep Singh9.6k
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1141 users visited in the last hour