gapseq fill ec-code error

[JosuaCarl]

Hello, when I am executing gapseq fill with my model, in the end I get:

Top 10 produced metabolites: NA:NA Error in .checkTypos(e, names_x) : Object 'ec' not found. Perhaps you intended seed Calls: addReactAttr ... tryCatchList -> tryCatchOne -> -> .checkTypos Execution halted

And when I print out mod@react_attr in the R file addReactAttr.R, I see, that it does not have a column named ec.

for bug reports and errors please report the output of: ./gapseq test Here is my gapseq test: gapseq: prediction and analysis of bacterial metabolic pathways and genome-scale networks

Usage: gapseq (find | find-transport | draft | fill | doall | adapt) (...) ... middle part omitted for shortness -h Show this screen. -n Enable noisy verbose mode.

I would be very glad to find out the cause of the error.

[jotech]

Hi @JosuaCarl this sounds strange indeed! I need a bit more details to see what is going wrong. Do you mind providing your draft model plus temporary files (rxnWeights.RDS, rxnXgenes.RDS) and the actual command you used for gapseq fill?

[JosuaCarl]

Hi @jotech So, I think I already got a little bit at what went wrong. I naively tried to use gapseq fill on my own model, which was drafted by CarveMe and used the naming convention from BiGG, while gapseq relies on Model SEED. I uses the rxnWeights and rxnXgenes, from gapseq find/find-transport/draft from the gene sequence of my organism. I suspect that the non matching sets of identifiers caused no ec-codes to be appended to the dataframe. The difference in naming convention unfortunatrly also makes it infeasible for me to compare both models, even when I use the completed draft by gapseq, but the method (gapseq fill) now works as it should. Thank you very much for your reply.

[jotech]

i see this makes sense and I agree the differences in namespace definitely caused the errors you encountered! For comparing bigg and seed namespaces, you might have a look at:

• https://github.com/jotech/gapseq/blob/master/dat/SEED2VMH_translation_edited.csv
• https://www.metanetx.org/mnxdoc/mnxref.html

which at least could at assist you in matching reactions ids. In addition, comparing model quality by using memote (https://memote.io) or the predicted phenotype vs. literature/databases (e.g. https://bacdive.dsmz.de/) could maybe another alternative?