V0.4

Project.toml

@@ -13,6 +13,7 @@ CodecZlib = "944b1d66-785c-5afd-91f1-9de20f533193"
 CodecZstd = "6b39b394-51ab-5f42-8807-6242bab2b4c2"
 DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"
 DelimitedFiles = "8bb1440f-4735-579b-a4ab-409b98df4dab"
+GeneticVariantBase = "2447270c-d849-4bf9-ac0d-b5c0b265991c"
 Glob = "c27321d9-0574-5035-807b-f59d2c89b15c"
 LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
 LoopVectorization = "bdcacae8-1622-11e9-2a5c-532679323890"

src/SnpArrays.jl

@@ -2,6 +2,7 @@ __precompile__()
 
 module SnpArrays
 
+using GeneticVariantBase
 using CodecZlib, CodecXz, CodecBzip2, CodecZstd,  TranscodingStreams
 using Adapt, Glob, LinearAlgebra, LoopVectorization, Missings, Mmap, Printf
 using Requires, SparseArrays, Statistics, StatsBase

src/iterator.jl

@@ -0,0 +1,105 @@
+mutable struct SnpArrayIterator <: VariantIterator
+   snpdata::SnpData
+end
+
+struct SnpArrayIndex <: Variant
+    index::Int
+end 
+
+@inline function Base.eltype(::Type{<:VariantIterator})
+    SnpArray
+end
+
+function Base.iterate(itr::SnpArrayIterator, state=itr.index)
+    if state > size(itr.snpdata.snp_info,2)
+        return nothing
+    else
+        state = state + 1
+        result = (view(itr.snpdata.snp_info, :, state), state+1)
+        return result
+    end
+end
+
+@inline function Base.length(itr::SnpArrayIterator)
+    return size(itr.snpdata.snp_info, 2)
+end
+
+@inline function snparrayiterator(snpdata::SnpData)
+    SnpArrayIterator(snpdata)
+end
+
+function iterator(snpdata::SnpData; startidx=1)::SnpArrayIterator
+    if startidx > size(snpdata,2)
+        return nothing
+    else
+        SnpArrayIndex(startidx)
+        return SnpArrayIterator(snpdata)
+    end
+end
+
+function iterator(s::SnpData; startidx=1)
+    iterator = SnpArrayIterator(s)
+end
+
+function chrom(s::SnpData, startidx=1)::String
+    result = view(s.snp_info,startidx,1)
+    return result
+end
+
+function pos(s::SnpData, startidx=1)::Int
+    result = view(s.snp_info,startidx,2)
+    return result
+end
+
+function rsid(s::SnpData, startidx=1)::String
+    result = view(s.snp_info,startidx,3)
+    return result
+end
+
+function alleles(s::SnpData, startidx=1)::Vector{String}
+    allele1 = view(s.snp_info,startidx,5)
+    allele2 = view(s.snp_info,startidx,6)
+    return [allele1, allele2]
+end
+
+function alt_allele(s::SnpData, startidx=1)::String
+    alt = view(s.snp_info,startidx,6)
+    return alt
+end
+
+function ref_allele(s::SnpData, startidx=1)::String
+    ref = view(s.snp_info,v,5)
+    return ref
+end
+
+function mafCount(s::SnpData,startidx=1)::Int
+    n00 = 0
+    n01 = 0
+    n11 = 0
+    for i in startidx:size(s,1)
+        row = s[i,:]
+        totalCount += 1
+        if row[5] == 0 && row[6] == 0
+            n00 += 1
+        else if row[5] == 0 && row[6] == 1
+            n01 += 1
+        else if row[5] == 1 && row[6] == 1
+            n11 += 1
+        end
+    end
+
+    return n00,n01,n11,totalCount
+end
+
+function maf(s::SnpData, startidx=1)
+    results = mafCount(s)
+    if results[1] <= results[2] && results[1] <= results[3]
+        return float(results[1]) / float(results[4])
+    end
+    if results[2] <= results[1] && results[2] <= results[1]
+        return float(results[2]) / float(results[4])
+    end
+    if results[3] <= results[1] && results[3] <= results[2]
+        return float(results[3]) / float(results[4])
+    end
+end

src/snpdata.jl

@@ -7,7 +7,7 @@ const PERSON_INFO_KEYS = [:fid, :iid, :father, :mother, :sex, :phenotype]
 
 Type to store SNP and person information along with the SnpArray.
 """
-struct SnpData
+struct SnpData <: GeneticData
     people::Int
     snps::Int
     snparray::SnpArray