Current status of monarchr

[bschilder]

Hi there,

I'm the developer of HPOExplorer (https://github.com/neurogenomics/HPOExplorer) and have been working on integrating various Monarch resources with HPO. I came across this package and thought it seemed very useful, but noticed there hasn't been activity for quite some time. I was wondering what the current status of this project was. Happy to make some PRs to get it up to date if needed (or perhaps it's working fine and none are needed!).

All the best,
Brian

[charlieccarey]

Hi Brian,

You're welcome to update it.

The project, as you suspected, has been quite dormant. I have no near term plans to do much more with it.

The github code and vignettes from 2018 were our most up to date code.

I just ran a couple simple queries. It has survived a bit better than I would have suspected.

As far as contributing, I suggest looking at https://github.com/monarch-initiative/monarchr. It is an official package from monarch initiative. Consider whether your long term efforts might be most lasting there? However that official package is a very basic package at this point and might serve a different purpose. Anyway, perhaps the author of that package might help you with other resources and contacts at monarch. In https://github.com/charlieccarey/monarchr I have strived to return simple useful objects along with sufficient examples and documentation. Continuing in that vein might make the effort to extend the current package worthwhile.

In the course of re-running the vignette disease_to_model_organism, (https://charlieccarey.github.io/monarchr/articles/disease_to_model_organism.html), I noticed:

1. Some of the functionality has broken. For instance there seems to be an error when selecting the target organisms for homologs. See issue bioentity_homologs might be ignoring homolog_taxon #11. This might be a useful place to try to dive into the monarch api documentation to see if there is a good fix from the query side. I honestly don't recall how I came up with homolog_taxon as a (previously?) valid query parameter. I suspect that in the response there was a column with that name or that I downloaded and inspected the json results while running queries manually through the 1st version of the web app. Or somewhere in the swagger documentation? Perhaps the API has changed, or is in the process of changing so that that column is no longer available. (Or might reappear in parallel with the additional features they hint are on the way with the newer version of the monarch web app?). In general, I had difficulty obtaining good documentation of the API, or understanding how best to do so.

2. Some of the results seem unexpectedly different vs. previous results. For example fewer organisms recovered with interactions after running the following line.

lapply(bap1_homolog_intxns_split, nrow)

I used to also recover interactions from rat and zebra fish here, which I currently am missing. I suspect that this is more a data representation problem or data ingest problem in monarch or their sources. Perhaps their interaction sources have changed. But I don't know. It is a little discouraging to no longer see the results for the other organisms. Hopefully it is something simple to work-around, or something they can fix on their side.

At any rate, your contributions would be welcome. In which case I'd be interested in seeing which directions you take it.

Best to you,

Chuck

[bschilder]

The github code and vignettes from 2018 were our most up to date code.

I just ran a couple simple queries. It has survived a bit better than I would have suspected.

I was also encouraged by that! a lot of API-dependant packages tend to break pretty quickly if not maintained.

As far as contributing, I suggest looking at https://github.com/monarch-initiative/monarchr. It is an official package from monarch initiative. Consider whether your long term efforts might be most lasting there? However that official package is a very basic package at this point and might serve a different purpose. Anyway, perhaps the author of that package might help you with other resources and contacts at monarch. In https://github.com/charlieccarey/monarchr I have strived to return simple useful objects along with sufficient examples and documentation. Continuing in that vein might make the effort to extend the current package worthwhile.

Thanks so much for letting me know about this, i hadn't come across it before somehow! I'll check it out and see if it fits my particular use case. I've decided to extract all the KG-related functions i've been working on and put them all in a new package called KGExplorer:
https://github.com/neurogenomics/KGExplorer

In the course of re-running the vignette disease_to_model_organism, (https://charlieccarey.github.io/monarchr/articles/disease_to_model_organism.html), I noticed:

1. Some of the functionality has broken. For instance there seems to be an error when selecting the target organisms for homologs. See issue bioentity_homologs might be ignoring homolog_taxon #11. This might be a useful place to try to dive into the monarch api documentation to see if there is a good fix from the query side. I honestly don't recall how I came up with homolog_taxon as a (previously?) valid query parameter. I suspect that in the response there was a column with that name or that I downloaded and inspected the json results while running queries manually through the 1st version of the web app. Or somewhere in the swagger documentation? Perhaps the API has changed, or is in the process of changing so that that column is no longer available. (Or might reappear in parallel with the additional features they hint are on the way with the newer version of the monarch web app?). In general, I had difficulty obtaining good documentation of the API, or understanding how best to do so.
2. Some of the results seem unexpectedly different vs. previous results. For example fewer organisms recovered with interactions after running the following line.

lapply(bap1_homolog_intxns_split, nrow)

I used to also recover interactions from rat and zebra fish here, which I currently am missing. I suspect that this is more a data representation problem or data ingest problem in monarch or their sources. Perhaps their interaction sources have changed. But I don't know. It is a little discouraging to no longer see the results for the other organisms. Hopefully it is something simple to work-around, or something they can fix on their side.

Interesting, the orthologous gene conversion is one of the core features of Monarch that I'm interested in. I have another package that focuses on this issue specifically, called orthogene. Though it doesn't yet use any Monarch resources, I'm considering integrating them into it if useful.
https://github.com/neurogenomics/orthogene

I'll look into this a bit more and let you know what I think the best way forward is.
Thanks again for the detailed response!