myScripts/interactionGeneNet.R at master · zzwch/myScripts · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
options(stringsAsFactors = F)
rm(list = ls())

# given source-target network
# output order of source and target for layout
getForceOrder <- function(net, iter = 100, seed = 666){
  source <- unique(net[,1])
  target <- unique(net[,2])
  nodes <- c(source, target)
  set.seed(seed)
  xy <- data.frame(x = sample(x = 1:length(nodes), size = length(nodes), replace = F),
                   y = 1, row.names = nodes)
  xy[target, 'y'] <- -1 # target_y = -1, source_y = 1

  kk <- (max(xy[,1]) - min(xy[,1]))/length(nodes)
  ka <- 100*kk
  kr <- kk
  max_disp <- 8
  # fa <- function(d, k){return(d^2/k)}
  # fr <- function(d, k){return(k^2/d)}
  mod <- function(d){abs(d[1])}
  for(i in 1:iter){
    # initial pos displacement
    u_disp = data.frame(x = rep(0, length(nodes)),
                        y = rep(0, length(nodes)),
                        row.names = nodes)
    # repulsive
    for(i in source){
      for(j in source){
        if(i == j) next()
        d <- xy[i,1] - xy[j,1]
        if(d == 0) d <- 1 # in case d==0
        u_disp[i,1] = u_disp[i,1] + (kr^2)/d
      }
    }
    for(i in target){
      for(j in target){
        if(i == j) next()
        d <- xy[i,1] - xy[j,1]
        if(d == 0) d <- 1 # in case d==0
        u_disp[i,1] = u_disp[i,1] + (kr^2)/d
      }
    }
    # attraction
    for(k in 1:nrow(net)){
      u <- net[k,1]
      v <- net[k,2]
      d <- xy[v,1] - xy[u,1]
      u_disp[u,1] = u_disp[u,1] + sign(d)*(d^2)/ka
      u_disp[v,1] = u_disp[v,1] - sign(d)*(d^2)/ka
    }

    for(l in nodes){
      u_disp_l <- mod(u_disp[l,])
      if(u_disp_l > max_disp)
        u_disp[l,1] <- u_disp[l,1]*max_disp/u_disp_l
      xy[l,1] = xy[l,1] + u_disp[l,1]
    }
  }
  #return(xy)
  return(list(source[order(xy[source,1])],
              target[order(xy[target,1])]))
}

# Plot source-target network as the top and bottom semocircles
circleInteractionBetween2CellType <- function(pair_inner_as_source,
                                              pair_inner_as_target,
                                              above_horizon_degree = c(5, 10),
                                              inner_r ,
                                              outer_r,
                                              size = 1, target_r = 0.01*size,
                                              nodes_order = c('force-directed','manual', 'random', "asis", "abc"),
                                              random_seed = 666,
                                              force_iter = 100,
                                              manual_order = list(up_inner = NULL,
                                                                  up_outer = NULL,
                                                                  dn_inner = NULL,
                                                                  dn_outer = NULL),
                                              link.curvature = 0.5, link.angle = 90, link.ncp = 5){
  # inner_type outer_type is charactor
  # pair_inner_as_* is data.frame with 2 columns (source, target)
  # nodes_order : asis is implemented, others to be done
  # I didnot consider  genes which can be ligand and receptor stimulatory
  nodes_order <- nodes_order[1]
  if(nodes_order == "random"){
    permutateRow <- function(x){
      x[sample(1:nrow(x), nrow(x), replace = F),]
    }
    set.seed(random_seed)
    pair_inner_as_target <- permutateRow(pair_inner_as_target)
    pair_inner_as_source <- permutateRow(pair_inner_as_source)
  }
  if(nodes_order == "abc"){
    pair_inner_as_target <- pair_inner_as_target[order(pair_inner_as_target[,2], pair_inner_as_target[,1]),]
    pair_inner_as_source <- pair_inner_as_source[order(pair_inner_as_source[,1], pair_inner_as_source[,2]),]
  }
  up_inner <- unique(pair_inner_as_target[,2])
  up_outer <- unique(pair_inner_as_target[,1])
  dn_inner <- unique(pair_inner_as_source[,1])
  dn_outer <- unique(pair_inner_as_source[,2])
  if(nodes_order == 'manual'){
    up_inner <- manual_order$up_inner
    up_outer <- manual_order$up_outer
    dn_inner <- manual_order$dn_inner
    dn_outer <- manual_order$dn_outer
  }

  if(nodes_order == 'force_directed'){
    tmp <- getForceOrder(net = pair_inner_as_target, iter = force_iter, seed = random_seed)
    up_inner <- tmp[[2]]
    up_outer <- tmp[[1]]
    tmp <- getForceOrder(net = pair_inner_as_source, iter = force_iter, seed = random_seed)
    dn_inner <- tmp[[1]]
    dn_outer <- tmp[[2]]
  }

  getXY <- function(items = NULL, from = 0, to = 180, r = 1){
    deg_iv <- seq(from = from*pi/180, to = to*pi/180, length.out = length(items))
    return(data.frame(items = items, x = r*cos(deg_iv), y = r*sin(deg_iv)))
  }

  xy_up_inner <- getXY(up_inner, above_horizon_degree[1], 180 - above_horizon_degree[1], r = inner_r)
  xy_dn_inner <- getXY(dn_inner, above_horizon_degree[1] + 180, 360 - above_horizon_degree[1], r = inner_r)
  xy_up_outer <- getXY(up_outer, above_horizon_degree[2], 180 - above_horizon_degree[2], r = outer_r)
  xy_dn_outer <- getXY(dn_outer, above_horizon_degree[2] + 180, 360 - above_horizon_degree[2],  r = outer_r)

  getLinks <- function(source_target, xy_source, xy_target, target_r){
    source_target <- as.data.frame(source_target)
    links = as.data.frame(cbind(
      xy_source[match(source_target[,1], xy_source[,1]), c(2, 3)],
      xy_target[match(source_target[,2], xy_target[,1]), c(2, 3)]
    ))
    colnames(links) <- c("x", "y", "xend", "yend")

    # shrink for arrow head
    xend <- links$xend + target_r*(links$x-links$xend)/sqrt((links$y-links$yend)^2 + (links$x-links$xend)^2)
    yend <- links$yend + target_r*(links$y-links$yend)/sqrt((links$y-links$yend)^2 + (links$x-links$xend)^2)
    links$xend <- xend
    links$yend <- yend
    return(links)
  }

  library(ggplot2)
  library(ggrepel)
  ggplot() +
    geom_curve(mapping = aes(x = x, y = y, xend = xend, yend = yend), color = "grey",
               data = getLinks(pair_inner_as_target, xy_up_outer, xy_up_inner, target_r = target_r),
               arrow = arrow(angle = 20, length = unit(0.1, "inches"),type = 'closed'),
               curvature = -link.curvature, angle = link.angle, ncp = link.ncp) +
    geom_point(mapping = aes(x, y), size = size, color = 'cyan', data = xy_up_outer) +
    geom_point(mapping = aes(x, y), size = size, color = "orange", data = xy_up_inner) +
    geom_text_repel(mapping = aes(x, y, label = items), data = xy_up_outer) +
    geom_text_repel(mapping = aes(x, y, label = items), data = xy_up_inner) +
    geom_curve(mapping = aes(x = x, y = y, xend = xend, yend = yend),  color = "grey",
               data = getLinks(pair_inner_as_source, xy_dn_inner, xy_dn_outer, target_r = target_r),
               arrow = arrow(angle = 20, length = unit(0.1, "inches"),type = 'closed'),
               curvature = -link.curvature, angle = link.angle, ncp = link.ncp) +
    geom_point(mapping = aes(x, y), size = size, color = "orange", data = xy_dn_inner) +
    geom_point(mapping = aes(x, y), size = size, color = 'cyan', data = xy_dn_outer) +
    geom_text_repel(mapping = aes(x, y, label = items), data = xy_dn_inner) +
    geom_text_repel(mapping = aes(x, y, label = items), data = xy_dn_outer) +
    geom_hline(yintercept = 0) +
    #    coord_fixed(ratio = 1) +
    theme_void()
}

# generage random source-target network for illustration
generate_net <- function(source, target, n_links = round(sqrt(length(source) * length(target))*3), seed = 666){
  set.seed(seed)
  return(unique(data.frame(source = sample(source, size = n_links, replace = T),
                           target = sample(target, size = n_links, replace = T))))
}
up_net <- generate_net(paste0("MC-S", 1:18), paste0("EC-T", 1:22), seed = 111)
dn_net <- generate_net(paste0("EC-S", 1:18), paste0("MC-T", 1:22), seed = 222)

circleInteractionBetween2CellType(inner_r = 1.2, outer_r = 2,  above_horizon_degree = c(10,10),
                                  pair_inner_as_source = up_net,
                                  pair_inner_as_target = dn_net,
                                  link.curvature = -0.1, link.angle = 30, link.ncp = 20,size = 6,
                                  nodes_order = 'random',  random_seed = 888)
circleInteractionBetween2CellType(inner_r = 1.2, outer_r = 2,  above_horizon_degree = c(10,10),
                                  pair_inner_as_source = up_net,
                                  pair_inner_as_target = dn_net,
                                  link.curvature = -0.1, link.angle = 30, link.ncp = 20,size = 6,
                                  nodes_order = 'force_directed',
                                  force_iter = 100,
                                  random_seed = 888)