I had sequenced four RNA samples on illumina (100bpPE). Two of them were control samples and two of them were mutant samples. Both the mutant samples showed 4-10% rRNA after alignment, but both the controls showed 56-60% rRNA. Instead of rRNA removal kits like Ribominus, we had used PolyA magnetic bead mRNA purification, and it was expected that rRNA would be removed. The experiment was also performed with some drug treated samples. The same problem reappeared. The control samples still have ample amount of rRNA. Can anybody suggest something to clean up the experiment? Or else, if it could be the nature of the control biological sample, please suggest some biological possibilities for such occurrences and probable reasons for the same. I want to use these results for differential expression of genes, and I fear that this would affect my analysis.
You don't mention what species this is, but I'm assuming polyA capture makes sense here and that you're not looking at bacteria.
To me, this looks like a problem with your library preparation. Ensure you treated both sets of samples in *exactly* the same way. Did use the same amount of input total RNA for both sets of samples? Did you make sure not to overload the polyA beads and run it at the correct temperature/time? Check your equipment/protocol carefully.
I recently heard of a very expensive ChIP-seq experiment failing because the lab PCR machine was faulty. It took them months to figure it out.
You are right, a systematic problem like this is going comprimise your experiment. You could try throwing away your rRNA reads, but as this artifact affects 50% your control reads it is likely to influence your results by more than the mutation.
Thank you Rohan and Christian for your suggestions. I can make some things clear. both the samples were grown in identical conditions, given identical treatments,same concentration of RNA was used as the starting material -1ug. But the bioanalyser results of the control showed low RIN number even though the graph showed good intact RNA. It was suggested that the may be the samples contains small RNAs.
The organism we work on is plant pathogen, a filamentous fungus. The reads which did not align to mRNA was aligned to rRNA against rRNA sequence from database. Thus we could say that about 60% of unaligned sequences were that of rRNA.
Is there any chance of finding polyadenylated rRNA ?
Yes, do you have any comments on my previous post ,about the polyadenylation of rRNA?
Thanks. Such discussions help a lot. I have no idea how u saw the data, if yes good. no need the explain. But how could you find same error in every time. Even the same is true for drug treated samples. The mutant samples doesn't show this high rRNA. Both the samples were extracted the same way and processed the same way together. Can you please explain why is this type difference in both the samples?. the bioanalyzer data was generated before and after polyA enrichment. we have not used Ribominus. The RNA was freshly prepared.
The service providers had already re sequenced both the control and treated samples but since they had this problem repeatedly, they did some normalisations before sending it to me.. I had gone through the nature ref you had send , it was helpful. It said that the rRNA removal may depend on the different methods of rRNA removal as well as the type of culture you handle. I currently want to look at specifically the mRNA, and that is why poly A purification was done. Spending so much money on a single experiment is not feasible as far as a university set up is concerned and it cannot go waste. I will have to rethink about ways to analyse this data.
Well, it seams to be a manipulation problems since you have your mutant with expected rRNA reads percentage. The most powerful techniques for rRNA cleaning for RNA-seq are Ribominus, Ribozero and you can try the 5'-terminator exonuclease (Life technologies). This enzyme is a monophosphate-dependent exonuclease that cut predominantly rRNA. bi or triphosphate 5' RNA (as non rRNA) are protected. You can also combine PolyA+5'-terminator.
Varsha, did the service providers regenerate the libraries or just do the resequence them. If they only resequenced the libraries, then I'm not surprised they got the same. I strongly believe this is an error in library prep. Get them to redo the libraries, if they haven't already.
Thank you all...I shall check with them whether they resequenced the libraries or the library was also remade. I shall also try using ribominus or ribozero if I can convince my PI to rerun the samples once again from scratch.
- Badges
- Science topic
- Similar topics
- Solid State Physics
- Devices