//Creates stats like average Orf length, Codon usage when using a random strand to compare to
public static Statistics calculateStatistics(NucleotideStrand experimentalStrand, DataTable dtAminoAcid)
{
NucleotideStrand strand = new NucleotideStrand();
strand.sequence = generateRandomDNA(experimentalStrand.length, experimentalStrand.length);
int OrfCount = 1;
strand.setLength();
List <Orf> OrfList = new List <Orf>();
//Reading Strand searching for Orf
for (int i = 0; i < (strand.length - 3); i++)
{
//Setting variables
string srchCode = strand.sequence.Substring(i, 3);
DataRow srchRow = dtAminoAcid.Rows.Find(srchCode);
Codon srchCodon = new Codon(srchRow[0].ToString(), srchRow[1].ToString(), Convert.ToDecimal(srchRow[2]));
//If start Codon is found the strand is translate until stop Codon reached
if ("Met" == srchCodon.name)
{
Codon currCodon = srchCodon;
int b = i;
Orf currOrf = new Orf();
currOrf.start = b;
Exon currExon = new Exon();
currExon.start = b;
while ((b < (strand.length - 12)) && (currCodon.name != "Stp"))
{
string currCode = strand.sequence.Substring(b, 3);
DataRow currRow = dtAminoAcid.Rows.Find(currCode);
currCodon = new Codon(currRow[0].ToString(), currRow[1].ToString(), Convert.ToDecimal(currRow[2]));
currOrf.CodonList.Add(currCodon);
currExon.CodonList.Add(currCodon);
//Detecting Intron
if ("AGGT" == strand.sequence.Substring(b + 1, 4))
{
//Finish Exon
currExon.finish = b;
currExon.length = currExon.finish - currExon.start;
currOrf.ExonList.Add(currExon);
//Starting Intron
Intron currIntron = new Intron();
currIntron.start = b;
//Reading Intron
string IntronStr = strand.sequence.Substring(b, 8);
IntronStr = MyMethods.buildBranchPoint(IntronStr);
while ((b < (strand.length - 8)) && (IntronStr == "01234567"))
{
IntronStr = strand.sequence.Substring(b, 8);
IntronStr = MyMethods.buildBranchPoint(IntronStr);
b++;
}
//Converts CGs at end of strand into pyrimidines for comparison because the end of the strand contains alway contains a chain pyrimidines
IntronStr = strand.sequence.Substring(b, 12);
IntronStr = MyMethods.buildIntronFinishSeq(IntronStr);
while ((b < (strand.length - 12)) && ("PPPPPPPPXPAG" != IntronStr))
{
IntronStr = strand.sequence.Substring(b, 12);
IntronStr = MyMethods.buildIntronFinishSeq(IntronStr);
b++;
}
b = b + 12;
//Finishing Intron
currIntron.finish = b;
currIntron.length = currIntron.finish - currIntron.start;
currIntron.sequence = strand.sequence.Substring(currIntron.start, currIntron.length);
currOrf.IntronList.Add(currIntron);
//Starting new Exon
currExon = new Exon();
currExon.start = b;
}
//When stop Codon is found Orf is assessed as to how probable it is that it is a gene
if ("Stp" == currCodon.name)
{
//Finish Exon
currExon.finish = b;
currExon.length = currExon.finish - currExon.start;
currOrf.ExonList.Add(currExon);
//Finish Orf
currOrf.finish = b;
currOrf.length = currOrf.finish - currOrf.start;
int bpCount = currOrf.ExonList.Sum(item => item.length);
currOrf.setAACodonUsage(dtAminoAcid);
currOrf.setAAwholeUsage(dtAminoAcid);
currOrf.ExonLength = currOrf.ExonList.Sum(Exon => Exon.length);
int currVal = currOrf.finish;
bool myCheck = OrfList.Any(Orf => Orf.finish == currOrf.finish);
if (myCheck == false)
{
OrfList.Add(currOrf);
}
OrfCount++;
}
b = b + 3;
}
}
}
Statistics myStatistics = new Statistics();
double meanExonLength = (OrfList.Sum(Orf => Orf.ExonLength)) / OrfList.Count;
double StanDevExonLength = OrfList.Sum(Orf => Math.Pow(Orf.ExonLength - meanExonLength, 2));
StanDevExonLength = Math.Sqrt((StanDevExonLength) / OrfList.Count);
myStatistics.meanExonLength = meanExonLength;
myStatistics.StanDevExonLength = StanDevExonLength;
myStatistics.rangeExonLength = myStatistics.meanExonLength + myStatistics.StanDevExonLength;
double meanWholeLength = (OrfList.Sum(Orf => Orf.length)) / OrfList.Count;
double StanDevWholeLength = OrfList.Sum(Orf => Math.Pow(Orf.length - meanWholeLength, 2));
StanDevWholeLength = Math.Sqrt((StanDevWholeLength) / OrfList.Count);
myStatistics.meanLength = meanWholeLength;
myStatistics.StanDevLength = StanDevWholeLength;
myStatistics.rangeLength = myStatistics.meanLength + myStatistics.StanDevLength;
double MeanAACodonUsage = Convert.ToDouble((OrfList.Sum(Orf => Orf.AACodonUsage)) / OrfList.Count);
double StanDevAACodonUsage = OrfList.Sum(Orf => Math.Pow(Convert.ToDouble(Orf.AACodonUsage) - MeanAACodonUsage, 2));
StanDevAACodonUsage = Math.Sqrt((StanDevAACodonUsage) / OrfList.Count);
myStatistics.MeanAACodonUsage = MeanAACodonUsage;
myStatistics.StanDevAACodonUsage = StanDevAACodonUsage;
myStatistics.RangeAACodonUsage = myStatistics.MeanAACodonUsage + myStatistics.StanDevAACodonUsage;
double MeanAAwholeUsage = Convert.ToDouble((OrfList.Sum(Orf => Orf.AAwholeUsage)) / OrfList.Count);
double StanDevAAwholeUsage = OrfList.Sum(Orf => Math.Pow(Convert.ToDouble(Orf.AAwholeUsage) - MeanAAwholeUsage, 2));
StanDevAAwholeUsage = Math.Sqrt((StanDevAAwholeUsage) / OrfList.Count);
myStatistics.MeanAAwholeUsage = MeanAAwholeUsage;
myStatistics.StanDevAAwholeUsage = StanDevAAwholeUsage;
myStatistics.RangeAAwholeUsage = myStatistics.MeanAAwholeUsage + myStatistics.StanDevAAwholeUsage;
return(myStatistics);
}
//Looks for Orfs
public static List <Orf> findOrf(NucleotideStrand strand, DataTable dtAminoAcid, Statistics currStat)
{
int OrfCount = 1;
strand.setLength();
List <Orf> OrfList = new List <Orf>();
//Reading Strand searching for Orf
for (int i = 0; i < (strand.length - 3); i++)
{
//Setting variables
//DNA is broken into 3 character strings because that is how the body reads DNA. A 3 segment strand is called Codon
string srchCode = strand.sequence.Substring(i, 3);
DataRow srchRow = dtAminoAcid.Rows.Find(srchCode);
Codon srchCodon = new Codon(srchRow[0].ToString(), srchRow[1].ToString(), Convert.ToDecimal(srchRow[2]));
//If start Codon is found the strand is translate until stop Codon reached
if ("Met" == srchCodon.name)
{
Codon currCodon = srchCodon;
//b just symbols the current position in the strand
int b = i;
Orf currOrf = new Orf();
currOrf.start = b;
Exon currExon = new Exon();
currExon.start = b;
//Orfs can be be broken down into Exons and Introns, Introns don't follow the same design pattern as Exons
//Therefore to properly assess a gene the Exons and intros have to be seperated out
while ((b < (strand.length - 12)) && (currCodon.name != "Stp"))
{
string currCode = strand.sequence.Substring(b, 3);
DataRow currRow = dtAminoAcid.Rows.Find(currCode);
currCodon = new Codon(currRow[0].ToString(), currRow[1].ToString(), Convert.ToDecimal(currRow[2]));
currOrf.CodonList.Add(currCodon);
currExon.CodonList.Add(currCodon);
//Detecting Intron (AGGT indicates the start of an Intron)
if ("AGGT" == strand.sequence.Substring(b + 1, 4))
{
//Finish Exon
currExon.finish = b;
currExon.length = currExon.finish - currExon.start;
currOrf.ExonList.Add(currExon);
//Starting Intron
Intron currIntron = new Intron();
currIntron.start = b;
//Reading Intron
//A charactertic of an Intron is that it has a branch point in the middle so before
//The Intron has ended the branch must be found
string IntronStr = strand.sequence.Substring(b, 8);
//buildBranchPoint just put the current Intron string into the format 01234567 if each character meets
// The criteria of an Intron
IntronStr = MyMethods.buildBranchPoint(IntronStr);
while ((b < (strand.length - 8)) && (IntronStr == "01234567"))
{
IntronStr = strand.sequence.Substring(b, 8);
IntronStr = MyMethods.buildBranchPoint(IntronStr);
b++;
}
//Converts CGs at end of strand into pyrimidines for comparison because the end of the strand contains alway contains a chain pyrimidines
IntronStr = strand.sequence.Substring(b, 12);
IntronStr = MyMethods.buildIntronFinishSeq(IntronStr);
while ((b < (strand.length - 12)) && ("PPPPPPPPXPAG" != IntronStr))
{
IntronStr = strand.sequence.Substring(b, 12);
IntronStr = MyMethods.buildIntronFinishSeq(IntronStr);
b++;
}
b = b + 12;
//Finishing Intron
currIntron.finish = b;
currIntron.length = currIntron.finish - currIntron.start;
currIntron.sequence = strand.sequence.Substring(currIntron.start, currIntron.length);
currOrf.IntronList.Add(currIntron);
//Starting new Exon
currExon = new Exon();
currExon.start = b;
}
//When stop Codon is found Orf is assessed as to how probable it is that it is a gene
if ("Stp" == currCodon.name)
{
//Finish Exon
currExon.finish = b;
currExon.length = currExon.finish - currExon.start;
currOrf.ExonList.Add(currExon);
//Finish Orf
currOrf.finish = b;
currOrf.length = currOrf.finish - currOrf.start;
currOrf.ExonLength = currOrf.ExonList.Sum(Exon => Exon.length);
int bpCount = currOrf.ExonList.Sum(item => item.length);
//This calculates the Codon usage relative to what would be expected in a living organism
currOrf.setAACodonUsage(dtAminoAcid);
currOrf.setAAwholeUsage(dtAminoAcid);
//if the current found Orf has properties greater than the statistical ranges it will be added to the the list of OrfS
if ((currOrf.ExonLength > currStat.rangeExonLength) && (Convert.ToDouble(currOrf.AACodonUsage) > currStat.RangeAACodonUsage) && (Convert.ToDouble(currOrf.AAwholeUsage) > currStat.RangeAAwholeUsage) && (currOrf.length > currStat.rangeLength))
{
bool myCheck = OrfList.Any(Orf => Orf.finish == currOrf.finish);
if (myCheck == false)
{
OrfList.Add(currOrf);
}
}
OrfCount++;
}
b = b + 3;
}
}
}
return(OrfList);
}
private void RunDetector(string inputFileName, string outputFileName, int PercentageStart, int PercentageFinish, string ProgramFiles)
{
string fileDirectory = ProgramFiles + @"\" + "AminoAcidProbTable.txt";
//Getting DNA Strand
string[] geneStrandArray = File.ReadAllLines(inputFileName);
//Getting Codon usage values
DataTable dtAminoAcid = MyMethods.buildAminoAcidDataTable(fileDirectory);
int start = (geneStrandArray.Length / 100) * PercentageStart;
int finish = (geneStrandArray.Length / 100) * PercentageFinish;
//Calculating strand length
decimal percentageCoverage = Convert.ToDecimal(PercentageFinish - PercentageStart);
int strandLength = Convert.ToInt32(((Convert.ToDecimal(geneStrandArray.Length) * percentageCoverage) / 100));
//Reading gene
List <Orf> OrfList = new List <Orf>();
int numberOfWindows = strandLength / 1000;
int lineCount = start;
int lineLimit = start + 1000;
//Strands are normally too long to be analysed in one go so are broken down into windows
Statistics myStatistics = new Statistics();
for (int windowCount = 0; windowCount < numberOfWindows; windowCount++)
{
NucleotideStrand strand = new NucleotideStrand();
strand.sequence = "";
while (lineCount < lineLimit && lineCount < geneStrandArray.Length && lineCount < (start + strandLength))
{
strand.sequence = strand.sequence + geneStrandArray[lineCount];
lineCount++;
}
strand.sequence = strand.sequence.Replace("\r\n", "");
strand.sequence = strand.sequence.ToUpper();
strand.length = strand.sequence.Length;
if (windowCount == 0)
{
MessageBox.Show("Strand approximately " + (strand.length * numberOfWindows).ToString() + "bp long");
//Calculating statistical ranges
myStatistics = MyMethods.calculateStatistics(strand, dtAminoAcid);
}
//Reading strand
OrfList.AddRange(MyMethods.findOrf(strand, dtAminoAcid, myStatistics));
//-100 so thee is a slight overlap with windows
lineCount = lineCount - 100;
lineLimit = lineLimit + 1000;
}
//Ordering the list of OrfS by AACodonUsage and writing the top values to file
OrfList = OrfList.OrderByDescending(Orf => Orf.AACodonUsage).ToList();
string foundGenes = "";
//Building string of found genes and writing to file only want the top 10
for (int x = 0; (x < OrfList.Count) && (x <= 9); x++)
{
foundGenes = foundGenes + "<" + OrfList[x].getWholeSequence() + "> \r\n \r\n";
}
using (StreamWriter file = new StreamWriter(outputFileName))
{
file.WriteLine(foundGenes);
}
pgStatus.Value = 100;
lblStatus.Content = "Finished!";
MessageBox.Show("Program Finished file created: " + outputFileName);
}
