Tutorial_of_scMLnet_0.2.0.Rmd

---
title: "Tutorial of scMLnet_0.2.0"
date: "Compiled: January 11, 2021"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Tutorial of scMLnet}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)
```

## Required packages

```{r eval=FALSE}
library(Seurat)
library(Matrix)
library(dplyr)
library(scMLnet)
```

## Input data

```{r eval=FALSE}
# import sample data
data <- readRDS("./example/data.rds")
#data <- data[,sample(1:ncol(data),5000,replace = F)]

# import sample annotation
BarCluTable <- read.table("./example/barcodetype.txt", header = T, sep = "\t", stringsAsFactors = F)
#BarCluTable <- BarCluTable[match(colnames(data),BarCluTable$Barcode),]
```

## Define the receiver cell and sender cell

```{r eval=FALSE}
types <- unique(BarCluTable$Cluster)
LigClu <- "B cells"      
RecClu <- "Secretory"    
```

## Set parameters

```{r eval=FALSE}
# The parameters for Filter and Normalization
Raw.data = TRUE
min.cells = 3
min.features = 200

# The parameters for FindMarkers of Seurat
min_pct = 0.05
logfc = 0.15
pval = 0.05

# The parameters for Differential expression analysis
abundant.cutoff = 0.01
all_mean <- rowMeans(data)
hist(log10(all_mean), breaks=100, main="", col="grey80",
     xlab=expression(Log[10]~"average count"))
abline(v=log10(abundant.cutoff), col="red", lwd=2, lty=2)
abs_logFC = 1

# output folder
workdir <- "./output"
```

## Load the prior databases

```{r eval=FALSE}
# user can input their own interaction DB which required two column: source and target
Databases <- readRDS("./prior_knowledge/hm_potential_matrix_RWR.rds")

LigRec.DB <- Databases$LigRec_matrix %>% as.matrix() %>% reshape2::melt() %>%
  .[.$value > 0,] %>%
  select(source=Var1, target=Var2, score=value) %>%
  mutate(source=as.vector(source), target=as.vector(target)) %>% as_tibble()

RecTF.DB <- Databases$RecTF_matrix %>% as.matrix() %>% reshape2::melt() %>%
  .[.$value > quantile(.$value,0.9),] %>%
  select(source=Var1, target=Var2, score=value) %>%
  mutate(source=as.vector(source), target=as.vector(target)) %>% as_tibble()

TFTG.DB <- Databases$TFTG_matrix %>% as.matrix() %>% reshape2::melt() %>%
  .[.$value > 0,] %>%
  select(source=Var1, target=Var2, score=value) %>%
  mutate(source=as.vector(source), target=as.vector(target)) %>% as_tibble()
```

## Construction of Multi-layer Signaling Networks

```{r eval=FALSE}
netList <- RunMLnet(data, BarCluTable, RecClu, LigClu, 
                    abundant.cutoff, abs_logFC,
                    Raw.data, min.cells, min.features,
                    pval, logfc, min_pct,
                    LigRec.DB, RecTF.DB, TFTG.DB)
saveRDS(netList,"./example/netList.rds")
```

## Save and Visualization of Multi-layer Signaling Networks

```{r eval=FALSE}
# Save but do not draw
DrawMLnet(netList,LigClu,RecClu,workdir,plotMLnet = F)

# Save and draw
PyHome <- "D:/Miniconda3/envs/R36/python.exe" #for Window
DrawMLnet(netList,LigClu,RecClu,workdir,PyHome,plotMLnet = T)
```