US 11,056,214 B2
Dual sample melting curve cluster and cost analysis
Yang Yang, Sunnyvale, CA (US); Bradley Scott Denney, Irvine, CA (US); and Attaullah Seikh, Irvine, CA (US)
Assigned to Canon U.S.A., Inc., Melville, NY (US)
Filed by Canon U.S. Life Sciences, Inc., Rockville, MD (US)
Filed on Jun. 23, 2017, as Appl. No. 15/631,794.
Claims priority of provisional application 62/353,615, filed on Jun. 23, 2016.
Prior Publication US 2018/0011969 A1, Jan. 11, 2018
Int. Cl. G16B 40/00 (2019.01); G06F 17/18 (2006.01); G01N 21/64 (2006.01); G01N 25/04 (2006.01); C12Q 1/6844 (2018.01); G16B 20/20 (2019.01); G16B 40/30 (2019.01)
CPC G16B 40/00 (2019.02) [C12Q 1/6844 (2013.01); G01N 21/6428 (2013.01); G01N 25/04 (2013.01); G06F 17/18 (2013.01); G16B 20/20 (2019.02); G16B 40/30 (2019.02); G01N 2021/6439 (2013.01)] 17 Claims
1. A method for determining cluster centers for curves associated with biological reactions on a device having at least two biological samples, the method comprising:
providing a device having at least two biological samples;
providing a thermal system in communication with the device, the thermal system controlling a temperature within the device causing a biological reaction in each of the biological samples, wherein the biological reaction is a thermal melt reaction;
providing a controller in communication with the device, the controller providing instructions for:
generating at least one curve indicative of the biological reaction in each of the biological samples, each curve representing a change in a physical parameter associated with the biological sample as a function of the temperature within the device;
calculating an optimal cluster number for the generated curves;
building at least one tuple containing cluster centers represented by the curves, wherein the number of cluster centers in each tuple equals the optimal cluster number and each optimal cluster number corresponds to identified tuples;
for each tuple, calculating a cost value indicative of curves outside of the tuple fitting into clusters associated with the cluster centers in the tuple; and
returning cluster centers in a tuple having a minimum cost value.