Intensity extraction from oligonucleotide clusters for base calling

Patent metadata

Field	Value
Publication number	US-11989265-B2
Application number	US-202217902630-A
Country	US
Kind code	B2
Filing date	Sep 2, 2022
Priority date	Jul 19, 2021
Publication date	May 21, 2024
Grant date	May 21, 2024

Abstract

The technology disclosed extracts intensities from sequencing images for base calling target clusters and attenuates spatial crosstalk from neighboring clusters. The technology disclosed accesses a particular section from a plurality of sections of an image output by a sensor, the particular section of the image including at least one pixel depicting intensity emission values from a target cluster and neighboring clusters located across the sensor, and convolves the particular section of the image with a corresponding convolution kernel in a plurality of convolution kernels, to generate a feature map comprising a plurality of feature values. The technology disclosed further assigns a corresponding feature value to the target cluster based on feature values in the plurality of feature values adjoining a center of the target cluster, and processes the corresponding feature value assigned to the target cluster, to base call the target cluster.

First claim

What is claimed is:

1. A computer-implemented method of base calling, the computer-implemented method including:
accessing, from an image of clusters of oligonucleotides output by a sensor, a particular section from a plurality of sections of the image, the particular section of the image including at least one pixel depicting intensity emission values from a target cluster of oligonucleotides and neighboring clusters of oligonucleotides located across the sensor;
convolving the particular section of the image with a corresponding convolution kernel in a plurality of convolution kernels, to generate a feature map comprising a plurality of feature values;
assigning a corresponding feature value to the target cluster of oligonucleotides based on feature values in the plurality of feature values adjoining a center of the target cluster of oligonucleotides; and
processing the corresponding feature value assigned to the target cluster of oligonucleotides, to base call by determining a nucleotide base for the target cluster of oligonucleotides.

2. The computer-implemented method of claim 1 , wherein the corresponding feature value assigned to the target cluster of oligonucleotides is a weighted feature value generated using interpolation based on the feature values adjoining the center of the target cluster of oligonucleotides.

3. The computer-implemented method of claim 1 , wherein assigning the corresponding feature value to the target cluster of oligonucleotides based on feature values adjoining the center of the target cluster of oligonucleotides further comprising:
selecting one or more features that are nearest to the center of the target cluster of oligonucleotides; and
interpolating one or more feature values corresponding to the selected one or more features, to generate the corresponding feature value.

4. The computer-implemented method of claim 3 , wherein interpolating the one or more feature values includes at least one of bilinear interpolation, bicubic interpolation, interpolation based on weighted area coverage, Lanczos interpolation, or interpolation based on Hanning window.

5. The computer-implemented method of claim 1 , wherein assigning the corresponding feature value to the target cluster of oligonucleotides comprises:
based on a coordinate location of the center of the target cluster of oligonucleotides relative to coordinate locations of adjacent pixels in the particular section of the image, assigning the corresponding feature value to the target cluster of oligonucleotides.

6. The computer-implemented method of claim 1 , wherein assigning the corresponding feature value to the target cluster of oligonucleotides comprises:
based on a coordinate location of the center of the target cluster of oligonucleotides relative to coordinate locations of adjacent features in the feature map, assigning the corresponding feature value to the target cluster of oligonucleotides.

7. The computer-implemented method of claim 1 , wherein assigning the corresponding feature value to the target cluster of oligonucleotides comprises:
based on one or more center-to-center distances associated with the target cluster of oligonucleotides, assigning the corresponding feature value to the target cluster of oligonucleotides.

8. The computer-implemented method of claim 7 , further comprising:
determining the one or more center-to-center distances associated with the target cluster of oligonucleotides, the one or more center-to-center distances including (i) a first center-to-center distance between the center of the target cluster of oligonucleotides and a center of a first pixel of the particular section of the image, wherein the target cluster of oligonucleotides is within the first pixel, and (ii) a second center-to-center distance between the center of the target cluster of oligonucleotides and a center of a second pixel of the particular section of the image, the second pixel neighboring the first pixel in the particular section of the image.

9. The computer-implemented method of claim 7 , further comprising:
determining the one or more center-to-center distances associated with the target cluster of oligonucleotides, the one or more center-to-center distances associated with the target cluster of oligonucleotides including (i) a first center-to-center distance between the center of the target cluster of oligonucleotides and a center of a first feature of a plurality of features in the feature map, and (ii) a second center-to-center distance between the center of the target cluster of oligonucleotides and a center of a second feature of the plurality of features, the second feature neighboring the first feature in the feature map.

10. The computer-implemented method of claim 1 , wherein the corresponding convolution kernel is trained by applying a cost function and evaluating errors in the feature values, then updating coefficients in the corresponding convolution kernel.

11. The computer-implemented method of claim 1 , further comprising:
training a base caller to generate and/or update coefficients of the corresponding convolution kernel.

12. The computer-implemented method of claim 11 , wherein training the base caller comprises:
training the base caller using at least one of least squares estimation, ordinary least squares, least-mean squares, and recursive least-squares to generate and/or update the coefficients.

13. The computer-implemented method of claim 11 , wherein training the base caller comprises:
training the base caller during a sequencing run, to update coefficients of the corresponding convolution kernel.

14. The computer-implemented method of claim 13 , wherein training the base caller during the sequencing run comprises:
training the base caller using data from one or more sequencing cycles of the sequencing run, to update the coefficients of the corresponding convolution kernel; and
using the updated coefficients of the corresponding convolution kernel for convolution of images generated during subsequent one or more sequencing cycles of the sequencing run.

15. The computer-implemented method of claim 13 , wherein training the base caller during the sequencing run comprises:
training the base caller using data from a first sequencing cycle of the sequencing run, to update the coefficients of the corresponding convolution kernel; and
using the updated coefficients of the corresponding convolution kernel for convolution of images generated during a second sequencing cycle and one or more subsequent sequencing cycles of the sequencing run.

16. The computer-implemented method of claim 1 , further comprising:
capturing the image using a point and shoot image capturing system, a line scan image capturing system, and/or one or more CMOS (complementary metal oxide semiconductor) sensors.

17. The computer-implemented method of claim 1 , wherein the particular section from the plurality of sections of the image is a first section that is generated from a first portion of a flow cell, wherein the corresponding convolution kernel is a first convolution kernel, the at least one pixel is a first pixel, the feature map is a first feature map, the plurality of feature values is a first plurality of feature values, the target cluster of oligonucleotides is a first target cluster of oligonucleotides, the corresponding feature value assigned to the target cluster of oligonucleotides is a first feature value, and wherein the computer-implemented method further comprises:
accessing a second section from the plurality of sections of the image output by a second

Assignees

• Illumina Inc

Inventors

• Kagalwalla Abde Ali Hunaid
• Ojard Eric Jon
• Mehio Rami
• Parnaby Gavin Derek
• Udpa Nitin
• Lu Bo
• Vieceli John S

Classifications

• G06F18/23213 Primary

  with fixed number of clusters, e.g. K-means clustering · CPC title

• G06F18/285

  Selection of pattern recognition techniques, e.g. of classifiers in a multi-classifier system · CPC title

• G06V10/56

  relating to colour · CPC title

• G06V10/454 Primary

  Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN] · CPC title

• G06V10/774

  Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting · CPC title