Fix charging for selenocysteine

[tahorst]

This makes some changes to the tRNA charging reactions flat files to get selenocysteine charging to work after the EcoCyc updates (#1065). To match the format of all other amino acids, our model uses a simplifying assumption to lump two reactions into one for Sel charging (charge with Ser and then convert Ser to Sel is replaced with direct charging with Sel). I've added back this one step reaction that does not match to a single EcoCyc reaction by adding a new file (trna_charging_reactions_added.tsv).

This is necessary because Sel charged tRNA steadily approaches 0 throughout the simulation. This difference in charged/uncharged levels will lead to much less ppGpp degradation leading to higher ppGpp concentrations.

Previous charging_molecules multigen plot (fraction tRNA charged steadily decreases for Sel, gold trace):

With updated charging reactions:

Other updates:

• Add a function to join data from multiple flat files when loading raw_data. This lets us separate manual data additions from automatically pulled EcoCyc files and can be applied to other flat files if desired.
• Simplify handling of raw_data.trna_charging_reactions by adding and removing data during raw_data load instead multiple locations in the parca.
• Remove empty column headers from raw_data loads - some EcoCyc files have an extra tab that led to empty columns being loaded. With the new data join function, there is a check to make sure columns match between files and this empty column was causing a difference.
• Remove tRNA charging reactions that are in trna_charging_reactions_removed.tsv but not in trna_charging_reactions.tsv

[ggsun]

Looks good to me! Thanks for addressing this issue, I hadn't realized selenocysteine charging would have a significant functional impact on the current model. Adding reactions directly into raw data also seems like a great decision.

[tahorst]

I hadn't realized selenocysteine charging would have a significant functional impact on the current model

It really only has an effect with ppGpp regulation which isn't enabled by default. SpoT senses the amount of uncharged tRNA to control the amount of ppGpp degradation so having extra uncharged tRNA around can have an effect. I don't have a good sense for how much it actually changes ppGpp concentrations but looked like about 10% less ppGpp degradation if Sel tRNA was fully charged vs fully uncharged.

Thanks for the typo fix!

[1fish2]

Might this PR fix #1084 in the anerobic nightly build?

[tahorst]

Might this PR fix #1084 in the anerobic nightly build?

There's a chance but I looked at the failure and I think it is because we don't get synthesis of an enzyme (ArgA) so Arg depletes. I think it relates more to regulation with ArgR and some other transcription factors.