Pseudogene in GEMs

[haowang-bioinfo]

There are 29 pseudogenes discovered in the HMR2 model based on the Ensembl annotation. Pseudogenes are identified with disruptions in ORF (frameshift, early stop codon). Given that most pseudogenes are neither transcribed nor translated, they should be removed from Genome-scale models that suppose to include only functional enzymes. I was wondering if similar cases have been mentioned in any publications or sysbio SOP @demilappa @mihai-sysbio?

It would be good to have more considerations and thorough discussions regarding this issue, which might be also relevant to the modeling work for other organisms. @BenjaSanchez @edkerk your comments are welcome.

[edkerk]

I'm unaware of any publications on this. I'm not sure how it is with human models though, these pseudogene identifications are based on a reference genome? Perhaps they are functional in some populations? Can you give an example of one of these pseudogenes from Ensembl? Meanwhile, I can't think of many situations where these kind of gene associations would really be problematic in GEM analysis?

[haowang-bioinfo]

The identification was based on Ensembl manual annotation, here is an example of pseudogene CDC14C.

[BenjaSanchez]

@edkerk @Hao-Chalmers I'm not a fan of adding too many pseudo-genes as it could lead to eventually biased results in gene esentiallity predictions (many of those genes could be non metabolic). However this is more relevant in single organisms for prediction i suppose

[haowang-bioinfo]

These pseudogenes in HMR2 have been manually checked and uploaded, together with other gene curation results, to the repo by commit 9d7a6ca. This issue has been resolved by this commit d9eb453.

[haowang-bioinfo]

Resolved.