breedHS/code/kinship.R at main · R-Computing-Lab/breedHS · 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
kinship <- function(ped){

  ## ped should be a matrix with 4 rows in the follwing order:
  ## ID, Father, Mother, Sex and should be ordered from from founding
  ## generation to current.  Sex may be coded as "M"/"F" or as 1/0.
  ##

  m.idx = ped[,4] == "M"
  f.idx = ped[,4] == "F"
  N = dim(ped)[1]

  if (sum(c(m.idx,f.idx)) == N){
    ped[m.idx,4] = 1
    ped[f.idx,4] = 0
  }else if (sum(c(m.idx,f.idx)) != 0) {
    sex.warning()
    return()
  } else if (sum(ped[,4] %in% c(0,1))<N){
    sex.warning()
    return()
  }

  ## Recode ids from 1:N ##h
  ped_ids <- ped[,1]
  ped.tmp = matrix(0,N,4)
  ped.tmp[,1] = 1:N
  for (i in 1:N){
    id.tmp = ped[i,1]
    idx = ped[,2]==id.tmp
    ped.tmp[idx,2] = i
    idx = ped[,3] == id.tmp
    ped.tmp[idx,3] = i
  }
  ped = ped.tmp

  ## Begin calculation ##
  kinship = diag(rep(1,N))
  rownames(kinship) <- ped_ids
  colnames(kinship) <- ped_ids

  for (i in 1:N){

    i.M = ped[i,2]
    i.F = ped[i,3]

    for (j in i:1){

      if (j==i){
        ## if j's parents are in the pedigree ##
        if (i.M!=0 & i.F!=0){
          kinship[i,j] = 1 + .5*kinship[i.M,i.F]
        }
      } else {
        if (i.M!=0 & i.F!=0){
          kinship[i,j] = .5*kinship[i.M,j] + .5*kinship[i.F,j]
          kinship[j,i] = kinship[i,j]
        }
      }

    }
  }
  return(kinship)
}


kinship.update<-function(ped.old, k.ped.old, ped.new){

  #m.idx = ped.new[,4] == "M"
  #f.idx = ped.new[,4] == "F"
  N.old = dim(ped.old)[1]
  N.new = dim(ped.new)[1]
  N = N.old + N.new
  #if (sum(c(m.idx,f.idx)) == N.new){
  #  ped.new[m.idx,4] = 1
  #  ped.new[f.idx,4] = 0
  #}else if (sum(c(m.idx,f.idx)) != 0) {
  #  sex.warning()
  #  return()
  #} else if (sum(ped.new[,4] %in% c(0,1))<N.new){
  #  sex.warning()
  #  return()
  #}

  ped = rbind(ped.old,ped.new)

  ## Recode ids from 1:N ##
  ped.tmp = matrix(0,N,4)
  ped.tmp[,1] = 1:N
  for (i in 1:N){
    id.tmp = ped[i,1]
    idx = ped[,2]==id.tmp
    ped.tmp[idx,2] = i
    idx = ped[,3] == id.tmp
    ped.tmp[idx,3] = i
  }
  ped = ped.tmp

  ## Begin calculation ##
  kinship = matrix(0,N,N)
  kinship[1:N.old,1:N.old] = k.ped.old

  for (i in (N.old+1):N){

    i.M = ped[i,2]
    i.F = ped[i,3]

    for (j in i:1){

      if (j==i){
        ## if j's parents are in the pedigree ##
        if (i.M!=0 & i.F!=0){
          kinship[i,j] = 1 + .5*kinship[i.M,i.F]
        }
      } else {
        if (i.M!=0 & i.F!=0){
          kinship[i,j] = .5*kinship[i.M,j] + .5*kinship[i.F,j]
          kinship[j,i] = kinship[i,j]
        }
      }

    }
  }
  return(kinship)
}


## Calculate empirical kinship ##
kinship.emp <- function(geno.v1 , geno.v2, map, method = "simple"){

  similar = 2 - abs(geno.v1 - geno.v2)
  no.comps = sum(is.na(similar)==FALSE)
  sum.sim = sum(similar , na.rm=TRUE)
  kin.emp = sum.sim / (2 * no.comps)
  return(kin.emp)
}


sex.warning <- function(){
  cat("There were problems found in the sex coding. Coding should be M/F or 1/0.\n")
}