Snp detection by melt curve clustering

A method of identifying a sequence variation between nucleotide sequences, the method comprising:

• receiving a plurality of sets of data points, each set corresponding to a different sample that contains copies of a double stranded molecule of two nucleotide sequences, each data point of a set including a signal value and a temperature value for the sample where the temperature increases for each successive data point, wherein each set defines a melt curve;
• determining a melt region having a melt region start and a melt region end;
• at least one processor performing a first normalization of each melt curve by:
• modifying the data points of the melt curve so that data points within an end region have an average value of a first number, wherein the end region is a temperature range starting at the melt region end;
• and
• modifying the data points of the melt curve such that the data points in a start region have an average value of a second number, wherein the start region is a temperature range ending at the melt region start;
• for each melt curve, identifying a threshold temperature at which the melt curve crosses a threshold;
• calculating an average threshold temperature from the respective threshold temperatures;
• shifting each melt curve so that the melt curve crosses the threshold at the average threshold temperature;
• performing a second normalization of each melt curve includes:
• modifying the data points of the melt curve having a lower temperature than the average threshold temperature such that the data points in the start region have an average value of a third number;
• assigning each melt curve to a respective cluster, wherein the melt curves assigned to a same cluster have one or more similar properties in the melt region relative to melt curves in other clusters;
• and
• identifying at least a portion of the nucleotide sequences corresponding to at least one cluster as having a sequence variation relative to the nucleotide sequences of another cluster.

A method of identifying a sequence variation between nucleotide sequences, the method comprising:

• receiving a plurality of sets of data points, each set corresponding to a different sample that contains copies of a double stranded molecule of two nucleotide sequences, each data point of a set including a signal value and a temperature value for the sample where the temperature increases for each successive data point, wherein each set defines a melt curve;
• determining a melt region having a melt region start and a melt region end;
• assigning each melt curve to a respective cluster, wherein the melt curves assigned to a same cluster have one or more similar shape properties in the melt region relative to melt curves in other clusters;
• at least one processor selecting a cluster of melt curves;
• the at least one processor determining a melting temperature of each melt curve of the selected cluster;
• the at least one processor grouping the melt curves of the selected cluster into a plurality sub-clusters based on the respective melting temperatures;
• and
• identifying at least a portion of the nucleotide sequences corresponding to at least one sub-cluster as having a sequence variation relative to the nucleotide sequences of another sub-cluster.

A method of identifying a sequence variation between nucleotide sequences, the method comprising:

• receiving a plurality of sets of data points, each set corresponding to a different sample that contains copies of a double stranded molecule of two nucleotide sequences, each data point of a set including a signal value and a temperature value for the sample where the temperature increases for each successive data point, wherein each set defines a melt curve;
• determining a melt region having a melt region start and a melt region end;
• at least one processor assigning each melt curve to a respective cluster by analyzing shapes of the melt curves, wherein the melt curves assigned to a same cluster have one or more similar shape properties in the melt region relative to melt curves in other clusters, wherein analyzing shapes includes:
• for each melt curve:
• calculating N average values, each value the average of one of a plurality of continuous segments of the melt curve;
• defining the set of N average values as a point in N-dimensional space;
• fitting the N-dimensional points to K N-dimensional functions;
• identifying each N-dimensional point with one of the K N-dimensional functions;
• and
• grouping the melt curves associated with a same N-dimensional function into a same cluster;
• and
• identifying at least a portion of the nucleotide sequences corresponding to at least one cluster as having a sequence variation relative to the nucleotide sequences of another cluster.

A method of identifying a sequence variation between nucleotide sequences, the method comprising:

• receiving a plurality of sets of data points, each set corresponding to a different sample that contains copies of a double stranded molecule of two nucleotide sequences, each data point of a set including a signal value and a temperature value for the sample where the temperature increases for each successive data point, wherein each set defines a melt curve;
• determining a melt region having a melt region start and a melt region end;
• at least one processor taking a negative first derivative of each melt curve to determine respective melt peak curves;
• the at least one processor assigning each melt curve to a respective cluster, wherein the melt curves assigned to a same cluster have one or more similar properties for the melt peak curves in the melt region relative to melt curves in other clusters;
• and
• identifying at least a portion of the nucleotide sequences corresponding to at least one cluster as having a sequence variation relative to the nucleotide sequences of another cluster.