/// <summary>
/// Read data into the specified SnpCollection from the SnpCollection
/// file having the specified filename (which may include a path).
/// The specified CancellationToken can be used to abort the read.
/// The progress parameter will be updated by this method with a
/// value of 0-100 to reflect the percent progress of the read.
/// </summary>
/// <param name="snps">The SnpCollection to receive the read data.</param>
/// <param name="filename">The path and filename of the SnpCollection file.</param>
/// <param name="cancel">The CancellationToken that can be used to abort the read.</param>
/// <param name="progress">The progress parameter that will be updated to reflect
/// the percent progress of the read.</param>
public static void Read(SnpCollection snps, string filename, CancellationToken cancel, Progress progress) {
if (snps == null) throw new ArgumentNullException("The SnpCollection cannot be null.");
if (filename == null) throw new ArgumentNullException("filename cannot be null.");
if (String.IsNullOrWhiteSpace(filename)) throw new ArgumentOutOfRangeException("filename cannot be empty.");
using (StreamReader reader = new StreamReader(filename)) {
long length = 0;
if (progress != null) length = reader.BaseStream.Length;
string[] columns = new string[6];
string line;
while ((line = reader.ReadLine()) != null) {
cancel.ThrowIfCancellationRequested();
line.FastSplit(',', columns);
byte chr = Convert.ToByte(columns[2]);
if (chr == 0) chr = 23; // handles legacy use of 0 for X
Snp snp;
if ((!String.IsNullOrWhiteSpace(columns[3]) && Char.IsDigit(columns[3][0]))) {
// new format snp file
snp = new Snp(columns[0], chr, Convert.ToInt32(columns[3]), Convert.ToSingle(columns[4]), columns[1], columns[5]);
} else {
// old SnpMap format snp file
snp = new Snp(columns[0], chr, -1, -1, columns[1], columns[3]);
}
snps.Add(snp);
if (progress != null) progress.Set(reader.BaseStream.Position, length);
}
}
}
public static void UpdateCentimorgans(string[] args) {
SnpCollection snps = new SnpCollection(1, 23);
string filename = args[1];
SnpFile.ReadRutgers(snps, filename, new System.Threading.CancellationToken(), null);
for (int i = 1; i <= 23; i++) snps.Add(new Snp("fake" + i, (byte) i, 0, 0, null, null));
SnpCollection refSnps = new SnpCollection(1, 23);
SnpFile.Read(refSnps, "RefSnps.csv");
Snp prevSnp = null;
foreach(Snp snp in refSnps) {
Snp rutgerSnp = null;
if (snps.Contains(snp.RsId)) {
rutgerSnp = snps[snp.RsId];
//if ((rutgerSnp.Chromosome != snp.Chromosome) || (Math.Abs(rutgerSnp.Position-snp.Position) > 10)) {
// Console.WriteLine("mismatched pos for " + snp.RsId + " - " + snp.Chromosome + ":" + snp.Position + " vs. " + rutgerSnp.Chromosome + ":" + rutgerSnp.Position);
// snp.cM = snps.ExtrapolateCentiMorganPosition(snp.Chromosome, snp.Position);
//} else {
// snp.cM = snps[snp.RsId].cM;
//}
snp.Chromosome = rutgerSnp.Chromosome;
snp.Position = rutgerSnp.Position;
}
if ((rutgerSnp != null) && (rutgerSnp.cM > 0)) {
snp.cM = rutgerSnp.cM;
} else {
snp.cM = snps.ExtrapolateCentiMorganPosition(snp.Chromosome, snp.Position);
}
if ((prevSnp != null) && (prevSnp.Chromosome == snp.Chromosome)
&& (prevSnp.cM > snp.cM)) {
Console.WriteLine("cM out of sequence " + prevSnp.RsId + "-" + snp.RsId);
}
prevSnp = snp;
}
SnpFile.Write(refSnps, "RefSnps2.csv");
}
/// <summary>
/// Read data into the specified SnpCollection from the Rutgers SNP
/// map file having the specified filename (which may include a path).
/// The specified CancellationToken can be used to abort the read.
/// The progress parameter will be updated by this method with a
/// value of 0-100 to reflect the percent progress of the read.
/// </summary>
/// </summary>
/// <remarks>See http://compgen.rutgers.edu/RutgersMap/AffymetrixIllumina.aspx </remarks>
/// <param name="snps">The SnpCollection to receive the read data.</param>
/// <param name="filename">The path and filename of the Rutgers SNP map file.</param>
/// <param name="cancel">The CancellationToken that can be used to abort the read.</param>
/// <param name="progress">The progress parameter that will be updated to reflect
/// the percent progress of the read.</param>
public static void ReadRutgers(SnpCollection snps, string filename, CancellationToken cancel, Progress progress) {
if (snps == null) throw new ArgumentNullException("The SnpCollection cannot be null.");
if (filename == null) throw new ArgumentNullException("filename cannot be null.");
if (String.IsNullOrWhiteSpace(filename)) throw new ArgumentOutOfRangeException("filename cannot be empty.");
using (StreamReader reader = new StreamReader(filename)) {
long length = 0;
if (progress != null) length = reader.BaseStream.Length;
string[] columns = new string[4];
string line;
reader.ReadLine(); // skip header
while ((line = reader.ReadLine()) != null) {
cancel.ThrowIfCancellationRequested();
line.FastSplit(',', columns);
byte? chr = Snp.ChromosomeToByte(columns[1]);
if (chr.HasValue && (chr.Value >= 1) && (chr.Value <= 23)) {
float cM;
Snp snp;
if (float.TryParse(columns[3], out cM)) {
snp = new Snp(columns[0], chr.Value, Convert.ToInt32(columns[2]), cM, null, null);
} else {
snp = new Snp(columns[0], chr.Value, Convert.ToInt32(columns[2]));
}
snps.Add(snp);
}
if (progress != null) progress.Set(reader.BaseStream.Position, length);
}
}
}
/// <summary>
/// Gets an SnpCollection of the SNPs contained in this Genome.
/// </summary>
/// <returns>An SnpCollection.</returns>
protected SnpCollection GetSnps() {
lock (lockObj) {
if (this.snps == null) {
var looseSnps = this.genotypes.Keys.ToArray<Snp>();
//Array.Sort(looseSnps);
ushort minChr = ushort.MaxValue;
ushort maxChr = ushort.MinValue;
foreach (var snp in looseSnps) {
if (snp.Chromosome < minChr) minChr = snp.Chromosome;
if (snp.Chromosome > maxChr) maxChr = snp.Chromosome;
}
snps = new SnpCollection(minChr, maxChr);
foreach (var snp in looseSnps) snps.Add(snp);
}
return this.snps;
}
}
