deck_00.Rmd

---
title: "Introduction to <grn>mrgsolve</grn><BR>Welcome"
author: Kyle T. Baron
date: Metrum Research Group, LLC
output: 
  slidy_presentation:
    number_sections: true
    css: [styles/slidystyles.css,styles/colors.css]
  html_document:
    theme: united
    number_sections: true
    toc: true
    css: [styles/htmlstyles.css,styles/colors.css]
---



```{r, include = FALSE}
library(mrgsolve)
library(dplyr)
library(knitr)
library(lattice)
opts_chunk$set(comment = '.', fig.height = 5, fig.width = 9, message = FALSE)
tryit_file <- "workbook.Rmd"
```

```{r setup, echo = FALSE, message = FALSE}
mod <- mread_cache("pk1", modlib()) %>% 
  update(end = 192, delta = 0.2) %>% Req(CP)
data(exidata)
data <- filter(exidata, ID <= 10)
set.seed(1222)
```


---

```{r, echo = FALSE}
mod %>% ev(amt = 100) %>% mrgsim(end = 72) %>% plot()
```

---

```{r, echo = FALSE}
mod %>% ev(amt = 100, ii = 24, addl = 4) %>% mrgsim() %>% plot()
```

---

```{r, echo = FALSE}
mod <- mread_cache("popex", modlib()) %>% Req(DV)
data <- expand.ev(amt = 300, ii = 24, addl = 4, ID = seq(20))
mod %>% data_set(data) %>% mrgsim() %>% plot()
```

---

```{r, echo = FALSE}
hmod <- mrgsolve:::house() %>% Req(DV,RESP) %>% update(end = 244)
hdat <- mutate(data, time = 24)
mrgsim(hmod, data = hdat, omega = dmat(0.1,0.1,0.1,0.1)) %>% plot()
```

---

```{r, echo = FALSE}
idata <- data_frame(VC = c(20,60,180))
hmod %>% mrgsim(idata = idata,  end = 24*28, delta = 0.1,
                events = ev(amt = 100, ii = 24, addl = 28)) %>%
  plot(DV~.)
```


---

```{r, echo = FALSE}
mod <- mread_cache("conway", "model")
e <- ev(amt = 1, evid = 8, time = 180) + ev(amt = 0, evid = 8, time = 2.5*365)
idata <- data_frame(L_0 = c(1,3,5,7,10))
out <- mrgsim(mod, events = e, idata = idata, end = 6*365, delta = 1)
out@data <- mutate(out@data, time = time - 180)
plot(out, L+logV~(time/365), xlab = "Time (year)")
```


# About `mrgsolve` 

- `R` package for simulation from ODE-based models
    - Free, OpenSource, GitHub, CRAN
- Language
    - Models written in `C++` inside model specification format
    - General purpose solver: `ODEPACK` / `DLSODA` (`FORTRAN`)
    - Simulation workflow in `R`
- Hierarchical (population) simulation
    - `ID`, $\eta$, $\varepsilon$
- Integrated PK functionaility
    - Bolus, infusion, `F`, `ALAG`, `SS` etc, handled under the hood
    - 1- and 2-cmt PK models in closed-form
- Extensible using `R`, `C++`, `Rcpp`, `boost`, `RcppArmadillo`
- `R` is it's natural habitat

# `mrgsovle` started as `QSP` modeling tool
- Motivation: large bone/mineral homeostatsis model (CaBone)
- History using
    - Berkeley Madonna
    - WinBUGS
    - NONMEM (attempted)
- 2010: write `R` front end to `deSolve`
- 2012: write `C++` interface to `DLSODA`
- Develop dosing / event capability
- More recently, expose functionality provided by
    - `Rcpp` - vectors, matrices, functions, environments, random numbers 
    - `boost` - numerical tools in `C++`
    - users' own `C++` code (functions, data structures, classes)
- Translator from `SBML` to `mrgsolve` using `R` bindings to `libSBML`   


# Orientation

- https://github.com/mrgsolve/learn

- GitHub site: https://github.com/metrumresearchgroup/mrgsolve

- mrgsolve website: https://mrgsolve.github.io

- User Guide: https://mrgsolve.github.io/user_guide

# Please download the zipped repo again

- https://github.com/mrgsolve/learn