We use a duplex-specific nuclease from Evrogen (http://www.evrogen.com/products/DSN/DSN.shtml) to reduce the amount of rRNA and tRNA in our RNA sequencing libraries that should work for you. Essentially what we do is ligate adapters, do RT and PCR, and then treat 50 ng of the library with the DSN enzyme and repeat the PCR. What the procedure does is denatures the DNA, allows for a long period of hybridization (5 hrs) at a high temperature, and then digests dsDNA. DNAs that are very abundant are more likely to hybridize during this time and get digested. For us, the most abundant DNAs prior to the treatment come from rRNAs and tRNAs. The most abundant in your case will be from this particular miRNA. After the treatment, you have enriched for less abundant clones, like those you are most interested in, and then repeat the PCR.

by size selection. After doing the per for library preparation, you can select the desired range for your pcr product.

We use a duplex-specific nuclease from Evrogen (http://www.evrogen.com/products/DSN/DSN.shtml) to reduce the amount of rRNA and tRNA in our RNA sequencing libraries that should work for you. Essentially what we do is ligate adapters, do RT and PCR, and then treat 50 ng of the library with the DSN enzyme and repeat the PCR. What the procedure does is denatures the DNA, allows for a long period of hybridization (5 hrs) at a high temperature, and then digests dsDNA. DNAs that are very abundant are more likely to hybridize during this time and get digested. For us, the most abundant DNAs prior to the treatment come from rRNAs and tRNAs. The most abundant in your case will be from this particular miRNA. After the treatment, you have enriched for less abundant clones, like those you are most interested in, and then repeat the PCR.

This might be an option

http://supportres.illumina.com/documents/myillumina/7836bd3e-3358-4834-b2f7-80f80acb4e3f/dsn_normalization_sampleprep_application_note_15014673_c.pdf

One could try clustering the reads which are highly abundant, that miRNA in your case. I would use FASTQ/A collapser from the FASTX tools, it generates clusters. If you want to remove these high coverage RNA's then one can download all the miRNA sequences and map your all Illumina reads on this dataset. In the next step use the unmapped reads for further processing. I do this using Bowtie2.

If you want to remove this specific RNA, you might try using a biotinylated probe which will anneal with your abundant RNA. Then you could pull out the DNA:RNA complex with Streptavidin beads (magnetic or agarose). Your purified small RNA should be depleted in this specific RNA. I think some of the kits for removal of ribosomal RNA work in a similar way.

Mario, It looks to me that you actually want to minimize the abundant miRNA in a miRNA-specific experiment. This means that you have probably already done a size selection and therefore Matthew's approach might work. Alternatively, you can utilize the renaturation kinetics he describes and run the duplex over a hydroxylapatite column to bind the duplexes preferentially formed from the most abundant species. This is called 'normalizing' a cDNA library.

If I understand well, you're running miRNA profiles on Illumina platform and you have a specific miRNA that is too much abundant in your libraries.

Then, if you want ot physically remove this miRNA from your small RNA sample before library prep you may look for miRNA-specific probes, possibly coupled to magnetic beads as it is usually done by rRNA removal kits. But I'm not sure it can be done, or at least I don't know if these kits/probes are commercially available. Otherwise, if you've already run your libraries, you can filter reads coming from these unwanted miRNA before mapping your reads.

Size selection is best if you want to look at mature miRNAs. If you are looking for the pre-miRNA form or the primary form of the miRNA you would need to use a different approach. Though the mature form is the only functional form, there is evidence that the percentage of miRNAs in the pre or the mature forms is one way that miRNA expression is regulated.

One other option -I have not tried it yet- is to block amplification of your abundant miRNA during the enrichment PCR. Which means you could do the enrichment PCR in the presence of a terminated oligo that partially anneals to the adapter and partially to your targeted miRNA.

Hi Valerio! I'm doing this sequencing analysis with Professor Mario and I have already filtered the reads from this particular miRNA but, as an example, if I have 2 million reads per sample before filtering, I have only 80.000- 120.000 after filtering for this miRNA. And I don't know if it is enough for me to do my analysis... Thank you very much for the answers.

Hi Juliana, I've some experience with RNA-Seq, but I've not yet perfomed small RNA sequencing.

However, it looks quite unusual to have a so highly abundant miRNA which represents about 90-95 % of the total amount of your miRNAs. Are you working with human samples?

However, the question is: "did you "biologically" expect all this amount of miRNA in your samples, or it may be due to over-amplification bias?

In the former case, well, you may choose to filter out these unwanted reads and work with the remaining. Obviously, the remaining reads should be unique (in terms of mapping position), otherwise you may not have the needed coverage for this kind of analysis. There are some papers (such as that of Zev Williams, PNAS 2013) in which the authors have a very low amount of starting material, due to the very few copies of circulating miRNAs (from plasma), and a quite comparable number of reads (about 120-130.000).

In the latter case you should first understand the origin of this over-amplification bias to avoid highly-biased results from your data analysis. I hope this may help you.

Yes, I'm working with whole blood samples from human, collected on PaxGene tubes. I believe that this specific miRNA, miR-486, comes from red blood cells (http://www.ncbi.nlm.nih.gov/pubmed/22510765). But I have also found another paper showing that this miRNA could be a bias in Illumina platform, reaching 50X more expression than others miRNAs (http://www.ncbi.nlm.nih.gov/pubmed/23663360), but it's not very informative...

Now I'll try to remove this miRNA with oligos and beads with streptavidin. I hope this could help me.

I thank you all for helping me with the answers.

We just ran our first miRNAseq library and have the same miR-486 popping up. Have you tried any of the suggestions? Or have recommendations on analysis to get usable data?

Hi Elizabeth! I have tried to extract this miRNA from samples using RiboMinus and an specific oligo. Now I'll resequence the same samples, after this procedure. I don't have results yet...

Did you extract miRNAs from blood? I saw a few papers that have the same miRNA as the most abundant miRNA in samples, using Truseq protocol and another fluids, such as plasma and CSF. I don't know if it is a Illumina bias...

Bom dia Juliana! I work with Libby Grass from Duke (earlier message) and also Antonio Cavarlho from UNIFESP who and we are having the issues with mir-486 when running mirSeq on whole blood samples. Most groups attribute mir-451 to hemolysis, not mir-451. Any new progress on your end?

hahaha... Bom dia, Mark!

We didn't have progress with this issue yet. I didn't have success with the first mir-486 extraction but I did some modifications and now I have to try it again.

We sequenced other samples that were collected in different tubes, with a different mirna extraction kit and we found the same miRNA as the most abundant... 

To sum up: it's still a mistery for us.

Have you tried doing qPCR for miR-486 to see if the finding is real? Typo error - in my last message - should have been "mir-451 to hemolysis, not mir-486"

Hi Mark!

Unfortunately, we didn't try to do it yet, cause we're trying to validate some experiments... But our next step is to validate these results, with special attention to mir-486.

In your case, some progress?

Hi,

Do you have any update in the field of depletion of highly abundant miRNA in samples? Has anybody tried the method recommended by Matthew R Willmann in miRNA library prep? I will be very grateful for tips as I have very similar situation as Juliana had.

Hi Justyna!

I don't have any update, unfortunately... But I can say that this finding is real because I did qPCR and found a Ct = 13 (average) to mir-486-5p. 

Hi Juliana,

any progress in the targeted depletion of mir-486?

 Hi Juliana, miR-486 also show very high expression in the plasma sequencing data in my study. Have you done any further research on this? 

 Hello Yan, hello Sten!

I haven't made any progress on this, unfortunately. The only thing I could say is that this expression is true because I confirmed that by qPCR. 

I could observe that miR-486-5p is the top 1 expressed miRNA in different kinds of samples prepared by Illumina Truseq kit. For example: http://bmcgenomics.biomedcentral.com/articles/10.1186/1471-2164-14-319 (exosomes)

And: http://rnajournal.cshlp.org/content/early/2013/03/22/rna.036863.112.full.pdf (CSF)

I really don't know if it could be a bias, as mentioned by Huang et al, 2013...

Hi all, 

What is the origin of the samples? In my experinece, miR-486 is a myomiR and is found abundantly in muscle cells and tissues, but also in plasma as a product of muscle cell death.

We looked across many datasets using Oasis (https://dzne.oasis.de) and found that miR-486-5p is expressed preferentially in blood cells and serum. This finding is very consistent and in many other tissues and cell types counts for miR-486-5p are close to zero (e.g. brain). Similar observations are true for miR-486-3p.

In other words, the expression seems to be extremely high in blood/serum and is sometimes close to absent in other tissues, meaning that you do not always get high counts of miR-486-5p when you prepare libraries with the Truseq kit. Cannot finally say something about a potential bias however...

You can design specific stem loops primers before the adapter ligation procedure, this will prevent adapters from binding to 486 which will be excluded from you library.

Hi all,

If someone needs an easy to use and straight forward solution, please check this out: Preprint Erythropoietic miR-486-5p and miR-451a depletion from whole ...