Training data generation for artificial intelligence-based sequencing

1. A computer-implemented method of determining metadata about clusters on a tile of a flow cell, the computer-implemented method comprising:

accessing a set of images of the tile captured during a sequencing run and preliminary center coordinates of the clusters determined by a base caller;

for each image set, obtaining, from the base caller, a base call classifying, as one of four bases, origin subpixels that contain the preliminary center coordinates and

a predetermined neighborhood of contiguous subpixels that are successively contiguous to respective ones of the origin subpixels,

thereby producing a base call sequence for each of the origin subpixels and for each of the predetermined neighborhood of contiguous subpixels;

generating a cluster map that identifies the clusters as disjointed regions of contiguous subpixels that are successively contiguous to at least some of the respective ones of the origin subpixels and share a substantially matching base call sequence of the one of four bases with the at least some of the respective ones of the origin subpixels; and

storing the cluster map in memory for use determining shapes and sizes of the clusters.

2. The computer-implemented method of claim 1 , further including training a classifier based upon the cluster map, the classifier being a neural network-based template generator for processing input image data to generate a decay map, a ternary map, or a binary map, representing one or more properties of each of a plurality of clusters represented in the input image data for base calling by a neural network-based base caller.

3. The computer-implemented method of claim 2 , further including:

determining hyperlocated center coordinates of analytes by calculating centers of mass of the disjointed regions of the cluster map as an average of coordinates of respective contiguous subpixels forming the disjointed regions; and

storing the hyperlocated center coordinates of the clusters in the memory for use as ground truth for training the classifier.

4. The computer-implemented method of claim 1 , wherein the cluster map identifies boundaries between contiguous subpixels whose base call sequences do not substantially match.

5. The computer-implemented method of claim 1 , wherein generating the cluster map comprises identifying as background those subpixels that do not belong to any of the disjointed regions.

6. The computer-implemented method of claim 1 , wherein the cluster map identifies cluster boundary portions between two contiguous subpixels whose base call sequences do not substantially match.

7. A non-transitory computer readable memory including instructions that, when executed by at least one processor, cause a system to implement actions for determining metadata about clusters on a tile of a flow cell, the actions comprising:

accessing a set of images of the tile captured during a sequencing run and preliminary center coordinates of the clusters determined by a base caller;

for each image set, obtaining, from the base caller, a base call classifying, as one of four bases, origin subpixels that contain the preliminary center coordinates and

a predetermined neighborhood of contiguous subpixels that are successively contiguous to respective ones of the origin subpixels,

thereby producing a base call sequence for each of the origin subpixels and for each of the predetermined neighborhood of contiguous subpixels;

generating a cluster map that identifies the clusters as disjointed regions of contiguous subpixels that are successively contiguous to at least some of the respective ones of the origin subpixels and share a substantially matching base call sequence of the one of four bases with the at least some of the respective ones of the origin subpixels; and

storing the cluster map in memory for use determining shapes and sizes of the clusters.

8. The non-transitory computer readable memory of claim 7 , further including instructions that, when executed by the at least one processor, cause the system to implement actions of training a classifier based upon the cluster map, the classifier being a neural network-based template generator for processing input image data to generate a decay map, a ternary map, or a binary map, representing one or more properties of each of a plurality of clusters represented in the input image data for base calling by a neural network-based base caller.

9. The non-transitory computer readable memory of claim 6 , further including instructions that, when executed by the at least one processor, cause the system to implement actions of:

determining hyperlocated center coordinates of the analytes by calculating centers of mass of the disjointed regions of the cluster map as an average of coordinates of respective contiguous subpixels forming the disjointed regions; and

storing the hyperlocated center coordinates of the clusters in the memory for use as ground truth for training the classifier.

10. The non-transitory computer readable memory of claim 5 , wherein the cluster map identifies boundaries between contiguous subpixels whose base call sequences do not substantially match.

11. The non-transitory computer readable memory of claim 7 , further including instructions that, when executed by the at least one processor, cause the system to implement actions comprising generating the cluster map by identifying as background those subpixels that do not belong to any of the disjointed regions.

12. The non-transitory computer readable memory of claim 7 , wherein the cluster map identifies cluster boundary portions between two contiguous subpixels whose base call sequences do not substantially match.

13. The non-transitory computer readable memory of claim 7 , further including instructions that, when executed by the at least one processor, cause the system to implement actions comprising generating the cluster map based on:

identifying origin subpixels at preliminary center coordinates of the clusters determined by the base caller; and

breadth-first searching for substantially matching base call sequences by beginning with the origin subpixels and continuing with successively contiguous non-origin subpixels.

14. A computer-implemented system including:

at least one processor;

a non-transitory computer readable memory including instructions that, when executed by at least a processor, cause the computer-implemented system to implement actions for determining metadata about clusters on a tile of a flow cell, the actions comprising:

accessing a set of images of the tile captured during a sequencing run and preliminary center coordinates of the clusters determined by a base caller;

for each image set, obtaining from the base caller, a base call classifying, as one of four bases, origin subpixels that contain the preliminary center coordinates and a predetermined neighborhood of contiguous subpixels that are successively contiguous to respective ones of the origin subpixels, thereby producing a base call sequence for each of the origin subpixels and for each of the predetermined neighborhood of contiguous subpixels;

generating a cluster map that identifies the clusters as disiointed regions of contiguous subpixels that are successively contiguous to at least some of the respective ones of the origin subpixels and share a substantially matching base call sequence of the one of four bases with the at least some of the respective ones of the origin subpixels; and

storing the cluster map in memory for use determining shapes and sizes of the clusters.

15. The computer-implemented system of claim 14 , further including instructions that, when executed by the at least one processor, cause the computer-implemented system to implement actions of training a classifier based upon the cluster map, the classifier being a neural network-based template generator for processing input image data to generate a decay map, a ternary map, or a binary map, representing one or more properties of each of a plurality of clusters represented in the input image data for base calling by a neural network-based base caller.

16. The computer-implemented system of claim 15 , further including instructions that, when executed by the at least one processor, cause the computer-implemented system to implement actions comprising:

determining hyperlocated center coordinates of analytes by calculating centers of mass of the disjointed regions of the cluster map as an average of coordinates of respective contiguous subpixels forming the disjointed regions; and

storing the hyperlocated center coordinates of the clusters in the memory for use as ground truth for training the classifier.

17. The computer-implemented system of claim 14 , wherein the cluster map identifies boundaries between contiguous subpixels whose base call sequences do not substantially match.

18. The computer-implemented system of claim 14 , further including instructions that, when executed by the at least one processor, cause the computer-implemented system to implement actions comprising generating the cluster map by identifying as background those subpixels that do not belong to any of the disjointed regions.

19. The computer-implemented system of claim 14 , wherein the cluster map identifies cluster boundary portions between two contiguous subpixels whose base call sequences do not substantially match.

20. The computer-implemented system of claim 14 , further including instructions that, when executed by the at least one processor, cause the computer-implemented system to implement actions comprising generating the cluster map based on:

identifying origin subpixels at preliminary center coordinates of the clusters determined by the base caller; and

breadth-first searching for substantially matching base call sequences by beginning with the origin subpixels and continuing with successively contiguous non-origin subpixels.