Duplicate SMILES cntd.

[simolon]

I understand that there are duplicate structures for tails 4 and 5 in the article figure and also saw duplicate smiles codes, due to this, in the data, as described here, #12 (comment)

Do you have a clean version of the experimental data where the values for the trans and cis lipids can be distinguished, and could share it? Or is there a way to derive it from the shared dataset, e.g. through the sequence of the tails you used when constructing the lipids library from the components?

Thanks a lot!

[cpuxuyue]

The SMILES strings were enumerated using MARVIN (https://chemaxon.com/marvin), which has inherent limitations and cannot distinguish between cis and trans isomers. Addressing this issue would require manual annotation to clean the data, a highly time-intensive process. However, since the transfection potency of the cis and trans tails was found to be highly comparable, such efforts are unlikely to significantly enhance data quality.

[simolon]

Thanks for the explanations! I had hoped that the experimental data had been recorded or processed in a structure along that of the heat map, i.e. categorized by the three components. Then one would only need to identify which tail was what isomer, from possibly a single entry in the lab journal. But I of course could only guess what is available in terms of data and documentation, so I was probably too optimistic.

In any way, thanks for sharing the code and data!