Issues with MFE values in MFE_distribution_Fig4a.csv

[jdratcliff]

Congratulations on a great preprint. In looking at the sequence:MFE value pairs in experiment_data/MFE_distribution_Fig4a.csv, I noted that either the values have been sorted incorrectly (and are thus mismatched) or there was an error in the MFE processing algorithm used during analysis. For example, sequence UAUAUUUUAGUAUAAGAUGUACAAAAGUUUUUGAUACUUUAAGGAAUAGUUUAAGUCUAUUAUAUAUAA is reported to have the lowest MFE value in the dataset with -498.7999878. The same sequence processed using RNAFold from ViennaRNA reports a value of -6.20. Similarly, CGUAGUAGAUUGCGGAUUAGCAAAGAUUGGACUUCAUUGCACUUUGUUCUCGCGCUGGAAAAGCUAAAUUUGAUCCUUAUCGGCCGCAGGCCUGGUUGCUUCGGCGUCGACGUUCCAUCCCGGAAGACCGCUGGUUCAGGUAAAUGUUGUAUUGUCUAGGACAUAUCAACCAGCACCGCAAAAGGCCUACACCUGCGAACGUACGCUCCUGAGGAGGCAU is reported to have a value of -0.1 versus an RNAFold result of -60.0. I suspect the MFE values for all of the sequences are wrong.

Figure 4 in the preprint clearly uses the distributions reported in this .csv file, so it will have to be updated.

[ekkkkki]

Thank you for your observation. You are absolutely right, that we had an misalignment where sequences and MFEs in MFE_distribution_Fig4a.csv seemed to be sourced from different iterations when repeating the same experiment.

Actually upon recalculating MFEs of the sequences listed in MFE_distribution_Fig4a.csv, we can also find the same conclusion:

No statistical significance between natural and generated group (p-value = 0.888, two-sided Mann-Whitney U test)

I will fix the misalignment issue in MFE_distribution_Fig4a.csv shortly, and we are considering updating the plot in the next revision.