Viterbi decoder for microarray signal processing

Abstract

A system and method for region-based calling utilizes a probability distribution of a phi-transformed logarithmic ratio to determine a set of possible transition paths through markers and marker states, constructs a local evidence matrix for each of the markers and generates a total per-marker value for each segment in a discrete region.

1 - 10 . (canceled) 
     
     
         11 . A system for making region based-calls, comprising:
 a region, markers, a marker, an expected responsiveness, a logarithmic ratio, a phi-transformed logarithmic ratio, possible transition paths, a local path window, a local evidence matrix, a Viterbi decoder, a first segmenter, a second segmenter, a transition matrix, a state initial probability vector, a state vector, a segmentation finalizer, a user interface, a probability distribution, a sensor, and an array.   
     
     
         12 . The system according to  claim 11 , wherein the sensor is for transmitting a signal containing data regarding the region from the sensor, which reads an emission from the array. 
     
     
         13 . The system according to  claim 11 , wherein the logarithmic ratio is divided by the expected responsiveness to generate the phi-transformed logarithmic ratio. 
     
     
         14 . The system according to  claim 11 , wherein the probability distribution is of the phi-transformed logarithmic ratio for each state. 
     
     
         15 . The system according to  claim 14 , wherein the probability distribution is modeled by calculating an emission probability for each of the states. 
     
     
         16 . The system according to  claim 11 , wherein the marker is selected from the markers. 
     
     
         17 . The system according to  claim 11 , wherein the local path window is translated across the marker to determine the possible transition paths. 
     
     
         18 . The system according to  claim 11 , wherein the possible transition paths construct the local evidence matrix for the marker. 
     
     
         19 . The system according to  claim 11 , wherein the Viterbi decoder is operated with the transition matrix, the local evidence matrix and the state initial probability vector to generate the state vector. 
     
     
         20 . The system according to  claim 11 , wherein the state vector comprises a state associated with each of the markers along a most probable path through the local evidence matrix. 
     
     
         21 . The system according to  claim 11 , wherein the state vector operates the first segmenter to partition segments into a segmentation table of contiguous markers with the same state. 
     
     
         22 . The system according to  claim 21 , wherein the first segmenter operates the second segmenter with the segmentation table, and a minimum segment size to merge segments below the minimum segment size into adjoining segments to produce a minimized segmentation table. 
     
     
         23 . The system according to  claim 22 , wherein the second segmenter operates the segmentation finalizer on the minimized segmentation table to convert segment indexes to genomic positions, summarize the phi-transformed logarithmic ratio within each segment using the segment's median, and generate a total per-marker value for each segment. 
     
     
         24 . The system according to  claim 22 , wherein the segmentation finalizer is operated on the minimized segmentation table to convert segment indexes to genomic positions. 
     
     
         25 . The system according to  claim 23 , wherein the total per-marker value is displayed on the user interface. 
     
     
         26 . The system according to  claim 11 , wherein the system receives the region delineated by a chromosome, a start marker and an end marker. 
     
     
         27 . The system according to  claim 11 , wherein the local evidence matrix is constructed for each of the markers comprising a level of evidence for each of the states based on observed data. 
     
     
         28 . The system according to  claim 11 , wherein the phi-transformed logarithmic ratio within each segment uses a segment's median. 
     
     
         29 . The system according to  claim 21 , wherein a total per-marker value is generated for each segment. 
     
     
         30 . The system according to  claim 11 , wherein the system is configured for discrete region-based calling including receiving the region delineated by a chromosome, a start marker and an end marker and for each marker within the region, and dividing the logarithmic ratio for the marker by the expected responsiveness to copy number change to generate the phi-transformed logarithmic ratio.