---
title: "uberonpeek -- a look at UBERON ontology, etc., with ontoProc2"
author: "Vincent J. Carey, stvjc at channing.harvard.edu"
date: "`r format(Sys.time(), '%B %d, %Y')`"
vignette: >
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteIndexEntry{uberonpeek -- a look at UBERON ontology, etc., with ontoProc2}
  %\VignetteEncoding{UTF-8}
output:
  BiocStyle::html_document:
    highlight: pygments
    number_sections: yes
    theme: united
    toc: yes
---

# Introduction

The ontoProc2 package is designed to give convenient
access to the ontologies that are transformed to "semantic
SQL" in the INCAtools project.

We'll start by retrieving the current UBERON ontology and examining
some tables and "statements".

```{r setup,message=FALSE}
library(ontoProc2)
library(DBI)
library(dplyr)
ubss <- semsql_connect(ontology = "uberon")
report(ubss)
ubcon <- ubss@con
head(dbListTables(ubcon))
tbl(ubcon, "statements")
```

# Parent-child relations

CRAN's ontologyIndex package provides a familiar representation that
simplifies visualization.

```{r doconv, cache=FALSE}
uboi <- semsql_to_oi(ubcon)
uboi
uboi$name[10364:10370]
```

A sense of the variety of ontological cross-references
present can be given by tabling the tag prefixes.
```{r lktags}
prefs <- gsub(":.*", "", names(uboi$name))
table(prefs)
```

By using the ancestors component we can obtain
a view of is-a relations (presumably developed from
rdfs:subClassOf predicates).  We've chosen as
terminal tags the tags for heart, kidney, and cortex of kidney.
```{r doplot, fig.width=10}
onto_plot2(
  uboi,
  unlist(uboi$ancestors[c(
    "UBERON:0002189",
    "UBERON:0002113", "UBERON:0000948"
  )])
)
```

# EFO and NCI thesaurus

On cursory inspection, the EFO ontology has
considerable information about anatomic locations of diseases.

We'll use the entailed edges table in EFO to find all statements
that have 'heart' (UBERON:0000948) as object.

```{r lkncit}
eforef <- semsql_connect(ontology = "efo") # 240 MB
# nciref = semsql_connect("ncit")  # > 500MB, block
htab <- tbl(eforef@con, "entailed_edge") |>
  filter(object == "UBERON:0000948") |>
  as.data.frame()
head(htab)
```

It is tedious to see these formal tags.  We have
assembled a simple character vector map that covers
many tags.
```{r lkmap}
data(ncit_map)
head(ncit_map)
```

What are the predicates of the heart table above?
```{r lkpred}
ncit_map[unique(htab$predicate)]
```
To enumerate and decode the terms with disease location (EFO:0000784)
in heart, we have

```{r lkdec}
library(dplyr)
library(DT)
hdis <- ncit_map[unlist(htab |> dplyr::filter(predicate == "EFO:0000784")
  |> dplyr::select(subject))]
datatable(data.frame(tag = names(hdis), value = as.character(hdis)))
```

# Session information

```{r lksess}
sessionInfo()
```