QC reliability

[AMA-cs]

Hello,

I've been using shapemapper2 for processing my samples. My samples passed all the QC checks except one that always fail (Number highly reactive check):

As you can see in the screenshot, all the other checks are good. For high background check, the background is very low, which we saw in previous samples, and that's why we increased the reagent 1M7 for this sample hoping to pass the last check and get higher number of highly reactive nucleotides. Unfortunately, it didn't pass the check. I discussed the possible causes with the lab staff, who are running the experiments, but they eliminated most of them.

I'm wondering how reliable this check is, or how generalized it can be to assess the samples considering the different types of RNA? Also, our initial plan is to keep increasing the reagent concentration until we get enough reactivity while monitoring the background check. I'm looking for any insight/recommendations/suggestions to figure out how to improve our data quality experimentally or computationally.

Thank you

[Psirving]

90% of nucleotides have a mutation rates above background. This suggests that your 1M7 is working fine. I suspect the last possible cause: extremely highly structured RNA. Could you attach your reactivity profile pdf?

[AMA-cs]

@Psirving Thank you so much for your quick response. Attached is the pdf file for this specific sample. mRNA_profiles.pdf

However, for this QC. It failed with RNAs that we think they are not really highly structures. That's why I'm puzzled.

Thank you

[Psirving]

This is not an exceptionally highly structured RNA. It is likely failing because your reactivity rates are low.

If these are in-cell modifications, it could be that the reagent is not penetrating very well. It could also mean that your 1M7 is impure or partially degraded. 2A3 and NAI are reagents that produce reliably higher mutation rates and permeate cells better than 1M7. 2A3 is the gold standard, and is a very simple synthesis. NAI is commercially available through MilliporeSigma.

You could try again with one of these reagents, or you could proceed. This data looks sufficiently high quality to produce good structure models.

[AMA-cs]

@Psirving This is extremely helpful. Thank you so much. Our data is not in-cell. According to the lab person cell-free is the most relevant. I'm wondering if any of the points you mentioned still apply in this case?

Also, since our data is not in-cell, do you think if we keep increasing the reagent concentration, we might get higher reactivity rates eventually?

I think I will proceed with the current batch of data to generate structure models.

Thanks again

[Psirving]

In that case, it is likely that the effective concentration of 1M7 is lower than you think it is, either because of impurity or degradation.

NAI and 2A3 excel at in-cell modifications, but they are also a big improvement for extracted RNA. One reason that NAI and 2A3 are more reactive is their solubility. They are typically used at 100 mM final concentration. 1M7 is typically used at 10 mM final and will precipitate if you try to go much higher than that.

A third SHAPE reagent that I forgot to mention is 5NIA. It's readily available, cheap, and more shelf stable. Its not as good as 2A3 and NAI, but it is much more reactive than 1M7.