Methods, systems, and computer-readable media for accelerated base calling

• US 10,676,787 B2
• Filed: 10/12/2015
• Issued: 06/09/2020
• Est. Priority Date: 10/13/2014
• Status: Active Grant

• Current Assignee: Life Technologies Corporation (Thermo Fisher Scientific Incorporated)
• Original Assignee: Life Technologies Corporation (Thermo Fisher Scientific Incorporated)
• Inventors: Donnet, Benjamin
• Primary Examiner(s): Woitach, Joseph
• Application Number:
  US14/880,439
• Publication Number:
  US 20160102348A1
• Time in Patent Office: 1,702 Days
• Field of Search: None
• CPC Class Codes: C12Q 1/6869 Methods for sequencing
• CPC Class Codes: G16B 25/00 ICT specially adapted for hybridisation; ICT specially adapted for gene or protein expression

Abstract

Embodiments disclose methods, systems, and computer-readable media for accelerated base calling of sequencing data. These methods may be adapted to accelerate sequence determination for data arising from a variety of different nucleic acid sequencing platforms. In various embodiments, configurable logic circuits such as FPGAs and GPUs may be adapted to perform raw signal processing, basecalling, and/or sequence determination operations providing further enhancements to the sequence analysis methods.

Claims

1. A method for sample nucleotide sequence determination, the method comprising:

(a) flowing nucleotides onto a sensor array adapted for sequencing, the sensor array including a plurality of wells containing sample polynucleotides, wherein the nucleotides are flowed in a flow order;

(b) receiving from the sensor array, signal data relating to chemical reactions resulting from the flow of nucleotides, the signal data comprising a plurality of signal samples corresponding to the plurality of wells of the sensor array, wherein an input packet structure is configured to store the plurality of signal samples, the input packet structure including a bit field for each signal sample, a well identification bit field, a flow order identification bit field and a parameter bit field;

(c) accessing the plurality of signal samples from the input packet structure for a given well by a hardware-embedded basecaller, the hardware-embedded basecaller including a well buffer to store the plurality of signal samples for the given well and a plurality of parallel processing channels, each processing channel operable to test a particular nucleotide identification hypothesis that differs from each of the other processing channels, wherein each processing channel provides a putative basecall simultaneously with the other processing channels to produce a plurality of putative basecalls;

(d) selecting from the plurality of putative basecalls, a nucleotide identification basecall with a highest confidence of being attributable to a given nucleotide flow;

(e) repeating steps (c)-(d) for the signal samples acquired at each nucleotide flow of the flow order such that for each iteration, the basecaller extends an identified nucleotide sequence of the sample polynucleotide by at least one nucleotide identification basecall; and

(f) storing the identified nucleotide sequence for the sample polynucleotide in an output packet structure, the output packet structure including a sequence length bit field, a base bit field for each nucleotide identification basecall and a second well identification bit field.

2. The method of claim 1, wherein the plurality of parallel processing channels is configured to test the nucleotide identification hypotheses in groupings of four parallel processing channels, one for each nucleotide G, A, T, C of the flow order.

3. The method of claim 2, wherein the groupings of four parallel processing channels are further coupled to process sixteen nucleotide identification hypotheses in parallel, of which four putative basecalls are retained and processed in the next iteration by the basecaller.

4. The method of claim 1, wherein each processing channel applies a single-pass, non-iterative approximation of a gain and an offset to apply to the signal samples.

5. The method of claim 1, wherein the basecaller is implemented in a plurality of programmable electronic circuits.

6. The method of claim 1, wherein the basecaller is implemented in a graphics processing unit.

7. The method of claim 1, wherein the basecaller comprises a discrete component in a sequencing system, wherein the received signal data are generated by at least one sequencing instrument associated with the sequencing system, the method further comprising providing the identified nucleotide sequence for the sample polynucleotide to at least one data analysis server associated with the sequencing system.

8. The method of claim 1, wherein the receiving step, the input packet structure is received via a direct memory access (DMA) interface comprising a DMA receiver.

9. The method of claim 1, further comprising transmitting the output packet structure via a direct memory access (DMA) interface comprising a DMA transmitter.

10. The method of claim 1, wherein the basecaller is configured to operate using an independent clock.

11. The method of claim 1, wherein the basecaller is implemented in a field-programmable gate array.

12. The method of claim 4, wherein the output data packet further includes a gain bit field and an offset bit field, the method further comprising storing the gain in the gain bit field and the offset in the offset bit field of the output data packet.

13. The method of claim 1, wherein the output data packet further includes predicted data value bit fields, the method further comprising generating predicted data values by the basecaller and storing the predicted data values in the predicted data value bit fields of the output data packet.

14. The method of claim 1, wherein the output data packet further includes residual value bit fields, the method further comprising generating residual values by the basecaller and storing the residual values in the residual value bit fields of the output data packet.