/// <summary>
/// Implement the Maximum Contiguous Subsequence Sum alogirthm.
/// Source: http://www.cs.ucf.edu/~reinhard/classes/cop3503/lectures/AlgAnalysis04.pdf
/// Maximum Subarray Problem: http://en.wikipedia.org/wiki/Maximum_subarray_problem
///
/// Find the subsequence with the maximum sum of differences between the actual
/// base quality score and the given cutoff score.
/// </summary>
/// <param name="seqObj">The sequence object to be trimmed</param>
/// <returns>A new sequence object with trimmed sequence. Or null if a maximum subsequence
/// cannot be found (i.e. quality scores are below the cutoff)</returns>
public override ISequence Trim(ISequence seqObj)
{
byte[] scores = ((QualitativeSequence)seqObj).QualityScores.ToArray();
scores = QualityScoreAnalyzer.ConvertToPhred(scores, ((QualitativeSequence)seqObj).FormatType);
// Implement maximum sum segment algorithm.
int start = 0;
int sum = 0;
int maxSum = 0;
int maxStart = 0;
int maxEnd = -1;
for (int i = 0; i < scores.Length; i++)
{
sum += scores[i] - QualityThreshold;
// If sum is negative, the new subsequence resets from the next position.
if (sum < 0 && TrimFromStart)
{
start = i + 1;
sum = 0; // reset the sum to start the new subsequence
}
if (sum > maxSum)
{
maxSum = sum;
maxStart = start;
maxEnd = i;
}
}
// Return null if a maximum subsequence cannot be found
if (maxStart > maxEnd)
{
return(null);
}
var newLength = maxEnd - maxStart + 1;
// Also return null if the new trim length is less than the required minimum length
if (newLength < MinLength)
{
return(null);
}
var newSeqObj = seqObj.GetSubSequence(maxStart, newLength);
newSeqObj.ID = seqObj.ID;
return(newSeqObj);
}
