• Search
• Select
• Filter
• Look around/expand
• Plot

original doc: https://docs.google.com/document/d/1cbNP45X32G8BEoJgpX91gUANB5_CVCZARPh51RVgBA8/edit

This document is for rough notes on approaches for exposing Monarch via R

1. Fully featured R library with graph logic in R
2. thin wrapper onto new Monarch API
3. This wrapper onto OAK, which itself wraps Monarch API

Fully featured R library

The idea would be to have lots of ontology logic implemented in R. This would essentially be OAK in R.

I don’t think we should pursue this unless there is someone who has the time and resources, a strong vision and extensive experience of building similar libraries

Thin wrapper onto Monarch API

This would be a very lightweight layer that would mostly be geared around obtaining dataframes for genes, diseases, etc. It would have minimal logic. It could even be autoderived via swagger2R.

All logic would be server side. The R would mostly be retrieval

Thin wrapper onto OAK

This would require two new pieces
FastAPI layer onto OAK
An OAK adapter for the new Monarch API

However, both of these are already things we want.

There may be efficiency considerations with so many layers. There would need to be some thinking about the generic FastAPI layer for OAK (e.g. OAK is naturally sync; things like connecting to a sqlite for the first time incur an initial time overhead)

Trying to mass query all phenotypes in the KG by setting query to NULL and raising limit.

Running with limit<=500 works fine:

out <- monarchr::monarch_search(query = NULL,
                                      category = "biolink:PhenotypicFeature", 
                                      limit = 500)

But after some trial and error, I found out that anything >500 causes an error:

out <- monarchr::monarch_search(query = NULL,
                                      category = "biolink:PhenotypicFeature", 
                                      limit = 501)

Error in response_content$total : 
  $ operator is invalid for atomic vectors

Is this something caused by monarchR or the KG API itself?

Thanks!
Brian

Session info

R version 4.3.1 (2023-06-16)
Platform: aarch64-apple-darwin20 (64-bit)
Running under: macOS Sonoma 14.4.1

Matrix products: default
BLAS:   /System/Library/Frameworks/Accelerate.framework/Versions/A/Frameworks/vecLib.framework/Versions/A/libBLAS.dylib 
LAPACK: /Library/Frameworks/R.framework/Versions/4.3-arm64/Resources/lib/libRlapack.dylib;  LAPACK version 3.11.0

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

time zone: Europe/London
tzcode source: internal

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] monarchr_0.2.1

loaded via a namespace (and not attached):
 [1] vctrs_0.6.5       httr_1.4.7        cli_3.6.2         knitr_1.45        rlang_1.1.3       xfun_0.43         purrr_1.0.2      
 [8] generics_0.1.3    assertthat_0.2.1  jsonlite_1.8.8    glue_1.7.0        htmltools_0.5.8.1 fansi_1.0.6       rmarkdown_2.26   
[15] neo2R_2.4.2       evaluate_0.23     tibble_3.2.1      base64enc_0.1-3   fastmap_1.1.1     yaml_2.3.8        lifecycle_1.0.4  
[22] compiler_4.3.1    dplyr_1.1.4       igraph_2.0.3      pkgconfig_2.0.3   tidyr_1.3.1       rstudioapi_0.16.0 digest_0.6.35    
[29] R6_2.5.1          tidyselect_1.2.1  utf8_1.2.4        curl_5.2.1        pillar_1.9.0      magrittr_2.0.3    tidygraph_1.3.1  
[36] tools_4.3.1    

Hi there,

I'm in the process of developing another R package, KGExplorer, which seeks to merge knowledge graphs from multiple biomedical resources, with a special focus on Monarch to start. I'm quite partial to the tidygraph representation and plan to use that as one of the primary object classes, in addition to the ontology_DAG class within simona for ontologies.

Some potential integrations:

• neurogenomics/KGExplorer#1

I was also just having a discussion with @charlieccarey about their R package of the same name:

• charlieccarey/monarchr#10

@charlieccarey's package has some nice functionalities like searching for specific types of interactions, e.g. bioentity_interactions_assoc_w_gene

I'm mainly posting here to see if there's way we could coordinate our efforts. I haven't explored the package in your repo yet but would be happy to make contributions if there's room for it.

All the best,
Brian

Brian M. Schilder
PhD Candidate
Neurogenomics Lab
Faculty of Medicine, Department of Brain Sciences
UK Dementia Research Institute, Imperial College London
CV | bschilder.github.io/CV/CV
LinkedIn | linkedin.com/in/brian-schilder
Website I | bschilder.github.io/BMSchilder
Twitter | twitter.com/BMSchilder
Lab | neurogenomics.co.uk

Need monarch-initiative/monarch-app#240 for bulk search

w/ packagedown deployed to github pages

For checks + docs & site build & publish to GH pages.

Example:

eds_hits <- query_ids(c("MONDO:0007525", "MONDO:0020066", "MONDO:0034021"))
fetch_edges(eds_hits, result_categories = "biolink:Gene")

Warning: Unknown or uninitialised column: id.Error in tbl_kgx(nodes_df) :
Error: tbl_kgx nodes must have an 'id' column.