public void Generate(MBCO_association_class mbco_association, Ontology_type_enum ontology, params string[] bg_symbols)
{
this.Ontology = ontology;
this.Bg_genes = Array_class.Deep_copy_string_array(bg_symbols);
this.MBCO_association = mbco_association.Deep_copy();
if (this.Bg_genes.Length > 0)
{
MBCO_association.Keep_only_bg_symbols(bg_symbols);
}
}
private Ontology_enrichment_line_class[] Add_missing_process_information(Ontology_enrichment_line_class[] enrichment_lines)
{
enrichment_lines = enrichment_lines.OrderBy(l => l.Scp_name).ToArray();
int enrich_length = enrichment_lines.Length;
Ontology_enrichment_line_class enrich_line;
MBCO_association.Order_by_processName_symbol();
int onto_length = MBCO_association.MBCO_associations.Length;
int indexOnto = 0;
MBCO_association_line_class mbco_association_line;
int stringCompare;
for (int indexE = 0; indexE < enrich_length; indexE++)
{
enrich_line = enrichment_lines[indexE];
stringCompare = -2;
while ((indexOnto < onto_length) && (stringCompare < 0))
{
mbco_association_line = MBCO_association.MBCO_associations[indexOnto];
stringCompare = mbco_association_line.ProcessName.CompareTo(enrich_line.Scp_name);
if (stringCompare < 0)
{
indexOnto++;
}
else if (stringCompare == 0)
{
if (!string.IsNullOrEmpty(mbco_association_line.ProcessID))
{
enrich_line.Scp_id = (string)mbco_association_line.ProcessID.Clone();
}
enrich_line.ProcessLevel = mbco_association_line.ProcessLevel;
if (!string.IsNullOrEmpty(mbco_association_line.Parent_processName))
{
enrich_line.Parent_scp_name = (string)mbco_association_line.Parent_processName.Clone();
}
}
}
}
return(enrichment_lines);
}
private Ontology_enrichment_line_class[] Generate_enrichment_lines_and_calculate_pvalues(string[] experimental_symbols, Entry_type_enum entryType, int timepoint, string sample_name)
{
experimental_symbols = experimental_symbols.Distinct().OrderBy(l => l).ToArray();
int experimental_symbols_length = experimental_symbols.Length;
Dictionary <string, int> processName_symbol_count = new Dictionary <string, int>();
Dictionary <string, List <string> > processName_symbol_overlap_symbols = new Dictionary <string, List <string> >();
MBCO_association.Order_by_symbol_processName();
#region Count overlap between process genes and experimental genes
string experimental_symbol;
int mbco_associations_length = MBCO_association.MBCO_associations.Length;
int indexSymbol = 0;
int stringCompare;
MBCO_association_line_class mbco_association_line;
for (int indexMBCO = 0; indexMBCO < mbco_associations_length; indexMBCO++)
{
mbco_association_line = MBCO_association.MBCO_associations[indexMBCO];
stringCompare = -2;
while ((indexSymbol < experimental_symbols_length) && (stringCompare < 0))
{
experimental_symbol = experimental_symbols[indexSymbol];
stringCompare = experimental_symbol.CompareTo(mbco_association_line.Symbol);
if (stringCompare < 0)
{
indexSymbol++;
}
else if (stringCompare == 0)
{
#region Overlap symbols
if (!processName_symbol_overlap_symbols.ContainsKey(mbco_association_line.ProcessName))
{
processName_symbol_overlap_symbols.Add(mbco_association_line.ProcessName, new List <string>());
}
processName_symbol_overlap_symbols[mbco_association_line.ProcessName].Add(experimental_symbol);
#endregion
}
}
#region Process symbol count
if ((indexMBCO != 0) &&
(mbco_association_line.ProcessName.Equals(MBCO_association.MBCO_associations[indexMBCO - 1].ProcessName)) &&
(mbco_association_line.Symbol.Equals(MBCO_association.MBCO_associations[indexMBCO - 1].Symbol)))
{
throw new Exception();
}
if (!processName_symbol_count.ContainsKey(mbco_association_line.ProcessName))
{
processName_symbol_count.Add(mbco_association_line.ProcessName, 1);
}
else
{
processName_symbol_count[mbco_association_line.ProcessName]++;
}
#endregion
}
#endregion
#region Generate enrichment lines
string[] processNames = processName_symbol_overlap_symbols.Keys.ToArray();
string processName;
int processNames_length = processNames.Length;
Ontology_enrichment_line_class[] enrich_lines = new Ontology_enrichment_line_class[processNames_length];
Ontology_enrichment_line_class new_enrich_line;
for (int indexP = 0; indexP < processNames_length; indexP++)
{
processName = processNames[indexP];
new_enrich_line = new Ontology_enrichment_line_class();
new_enrich_line.Ontology_type = this.Ontology;
new_enrich_line.Scp_name = (string)processName.Clone();
new_enrich_line.Experimental_symbols_count = experimental_symbols_length;
new_enrich_line.Process_symbols_count = processName_symbol_count[processName];
new_enrich_line.Bg_symbol_count = this.Bg_genes.Length;
new_enrich_line.Overlap_symbols = processName_symbol_overlap_symbols[processName].OrderBy(l => l).ToArray();
new_enrich_line.Overlap_count = new_enrich_line.Overlap_symbols.Length;
new_enrich_line.EntryType = entryType;
new_enrich_line.Timepoint = timepoint;
new_enrich_line.Sample_name = (string)sample_name.Clone();
enrich_lines[indexP] = new_enrich_line;
}
#endregion
#region Calculate pvalues
int enrich_length = enrich_lines.Length;
Ontology_enrichment_line_class enrich_line;
int a; int b; int c; int d;
Fisher_exact_test_class fisher = new Fisher_exact_test_class(this.Bg_genes.Length, false);
for (int indexE = 0; indexE < enrich_length; indexE++)
{
enrich_line = enrich_lines[indexE];
a = enrich_line.Overlap_count;
b = enrich_line.Experimental_symbols_count - a;
c = enrich_line.Process_symbols_count - a;
d = enrich_line.Bg_symbol_count - a - b - c;
if ((a < 0) || (b < 0) || (c < 0) || (d < 0))
{
throw new Exception("negative values");
}
enrich_line.Pvalue = fisher.Get_rightTailed_p_value(a, b, c, d);
enrich_line.Minus_log10_pvalue = -(float)Math.Log10(enrich_line.Pvalue);
}
#endregion
#region Calculate qvalues and FDR
enrich_lines = enrich_lines.OrderBy(l => l.Pvalue).ToArray();
int enrich_lines_length = enrich_lines.Length;
int rank = 0;
for (int indexE = 0; indexE < enrich_lines_length; indexE++)
{
enrich_line = enrich_lines[indexE];
rank++;
enrich_line.Qvalue = enrich_line.Pvalue * ((double)processNames_length / (double)rank);
if (enrich_line.Qvalue > 1)
{
enrich_line.Qvalue = 1;
}
}
double lowest_qvalue = -1;
for (int indexE = enrich_lines_length - 1; indexE >= 0; indexE--)
{
enrich_line = enrich_lines[indexE];
if ((lowest_qvalue == -1) ||
(lowest_qvalue > enrich_line.Qvalue))
{
lowest_qvalue = enrich_line.Qvalue;
}
enrich_line.FDR = lowest_qvalue;
}
#endregion
return(enrich_lines);
}
public void Keep_only_indicated_level(int level)
{
MBCO_association.Keep_only_lines_with_indicated_level(level);
}
