Methods, systems, and computer readable media for making base calls in nucleic acid sequencing
A method for nucleic acid sequencing includes receiving a plurality of observed or measured signals indicative of a parameter observed or measured for a plurality of defined spaces; determining, for at least some of the defined spaces, whether the defined space comprises one or more sample nucleic acids; processing, for at least some of the defined spaces, the observed or measured signal to improve a quality of the observed or measured signal; generating, for at least some of the defined spaces, a set of candidate sequences of bases for the defined space using one or more metrics adapted to associate a score or penalty to the candidate sequences of bases; and selecting the candidate sequence leading to a highest score or a lowest penalty as corresponding to the correct sequence for the one or more sample nucleic acids in the defined space.
1 . A system, comprising:
a machine-readable memory; and
a processor configured to execute machine-readable instructions, which, when executed by the processor, cause the system to perform steps including:
for each of a plurality of defined spaces of a chip or substrate, wherein each defined space is in communication with a respective sensor and wherein at least some of the defined spaces comprise one or more sample nucleic acids, receiving, from the respective sensor for each defined space, an observed or measured signal indicative of a parameter observed or measured for the defined space in response to a nucleotide flow to the defined space, to provide a measured signal value representative of a non-incorporation or an incorporation of a base corresponding to the nucleotide flow by the sample nucleic acid in the defined space, wherein each defined space is a space in which at least some of a molecule, fluid, and/or a solid can be confined, retained and/or localized, wherein the plurality of defined spaces comprises at least 100,000 defined spaces;
generating, for the defined spaces comprising one or more sample nucleic acids, a set of candidate sequences of bases for the one or more sample nucleic acids of the defined space, including generating a data structure comprising a set of partial paths corresponding to candidate sequences undergoing an expansion in a stepwise manner one base at a time:
generating a set of predicted signal values for the set of candidate sequences based on a simulation framework for modeling responses by an active polymerase associated with the sample nucleic acid to a series of nucleotide flows, wherein the simulation framework includes possible states, state transitions and state transition parameters, wherein a first state of the possible states represents a non-incorporation of a particular base at a particular flow of the series of nucleotide flows and a second state of the possible states represents an incorporation of the particular base at the particular flow by the active polymerase associated with the sample nucleic acid;
applying one or more metrics adapted to associate a score or a penalty to each of the candidate sequences of bases, wherein at least one of the metrics depends on a residual between the measured signal value and the predicted signal value generated by the simulation framework, wherein the residual comprises a difference between the predicted signal value and the measured signal value; and
selecting from the set of candidate sequences of bases the candidate sequence
leading to a highest score or a lowest penalty as corresponding to a correct sequence for the one or more sample nucleic acids in the defined space.
2 . The system of claim 1 , wherein the step of generating the data structure comprises determining a distance between the predicted signal value for each partial path and the measured signal value.
3 . The system of claim 2 , wherein the one or more metrics comprise a metric that is a function of the distance between the predicted signal value for each partial path and the measured signal value.
4 . The system of claim 2 , wherein the one or more metrics comprise a metric comprising a sum of squared distances between corresponding values of the predicted signal value for each partial path and the measured signal value.
5 . The system of claim 1 , wherein the one or more metrics comprise a path metric comprising a sum of (i) a sum of squared residuals before an active window and (ii) a sum of squared residuals for negative residuals within the active window.
6 . The system of claim 1 , wherein the one or more metrics comprise a greedy decision metric comprising a sum of (i) a product of an empirical constant and a sum of squared residuals for negative residuals within an active window and (ii) a sum of squared residuals for positive residuals within the active window but only before an in-phase flow.
7 . The system of claim 1 , wherein the one or more metrics comprise a path metric, a greedy decision metric, and a per-flow metric, and wherein the per-flow metric comprises a weighted sum of (i) the path metric and (ii) the greedy decision metric.
8 . The system of claim 1 , wherein the one or more metrics comprise a scaled residual comprising a ratio between (i) a difference between a measured signal value for a current path at a current in-phase flow and a predicted signal value from a parent path at the current in-phase flow and (ii) a difference between a predicted signal value of the current path at the current in-phase flow and the predicted signal value from the parent path at the current in-phase flow.
9 . The system of claim 1 , wherein the one or more metrics comprise a total residual comprising a sum of squared residuals over all nucleotide flows.
10 . The system of claim 1 , wherein the step of generating the data structure comprises pruning the data structure using one or more absolute pruning rules selected from the group comprising: (i) discarding paths having a path metric larger than a best total residual metric, (ii) discarding paths having reached a last nucleotide flow, (iii) discarding paths for which a greedy decision metric exceeds a greedy penalty maximal threshold, (iv) discarding paths for which a polymerase activity is below a polymerase activity minimal threshold, (v) discarding paths including more than a threshold number of homopolymers having at least a threshold length in a row, (vi) discarding paths having a scaled residual metric below a first scaled residual minimal threshold, (vii) discarding paths having a scaled residual below a second scaled residual minimal threshold that is larger than the first scaled residual minimal threshold, and (viii) discarding paths having a greedy decision metric that exceeds the greedy decision metric for a best greedy expansion by more than a maximal threshold.
11 . The system of claim 1 , wherein the step of generating the data structure comprises pruning the data structure using one or more relative pruning rules selected from the group comprising: (i) discarding paths being more than a certain number of base pairs shorter than a longest path in the data structure, and (ii) discarding paths having a highest per-flow metric whenever the number of paths exceeds a certain threshold.
12 . The system of claim 1 , wherein the state transition parameters of the simulation framework comprise state transition parameters for the possible states of the active polymerase corresponding to a K-th base during an N-th nucleotide flow of the series of nucleotide flows, where K and N denote indices associated with the bases and the nucleotide flows.
13 . The system of claim 12 , wherein the simulation framework comprises first and second state transition parameters for the first and second states, respectively, for situations where the K-th base matches the N-th nucleotide flow, wherein (i) the first state transition parameter represents a proportion of the active polymerase that will remain active and not incorporate base K in flow N and (ii) the second state transition parameter represents a proportion of the active polymerase that will remain active and incorporate base K in flow N.
14 . The system of claim 13 , wherein the simulation framework further comprises:
determining the first state transition parameter by calculating a product of (i) a measure of a quantity of the active polymerase prior to a transition to the first state and (ii) a first transition factor [IER x(1−DR)]; and
determining the second state transition parameter by calculating a product of (i) the measure of the quantity of the active polymerase prior to a transition to the second state and (ii) a second transition factor [(1−IER)×(1−DR)],
where IER represents an incomplete extension rate and DR represents a droop rate.
15 . The system of claim 12 , wherein the simulation framework comprises first and second state transition parameters for the first and second states, respectively, for situations where the K-th base does not match the N-th flow, wherein (i) the first state transition parameter represents a proportion of the active polymerase that will remain active and not incorporate base K in flow N and (ii) the second state transition parameter represents a proportion of the active polymerase that will remain active and incorporate base K in flow N.
16 . The system of claim 15 , wherein the simulation framework further comprises:
determining the first state transition parameter by calculating a product of (i) a measure of a quantity of the active polymerase prior to a transition to the first state and (ii) a first transition factor [(1−CFR1)+(CFR″×IER×(1−DR))]; and
determining the second state transition parameter by calculating a product of (i) the measure of the quantity of the active polymerase prior to a transition to the second state and (ii) a second transition factor [CFR M ×(1−IER)×(1−DR)],
where IER represents an incomplete extension rate, DR represents a droop rate, CFR represents a carry forward rate, and M is a smallest number such that an (N−M) flow matches the K-th base.
17 . A method for nucleic acid sequence identification, comprising:
obtaining information corresponding to a plurality of observed or measured nucleotide incorporation events for one or more sample nucleic acid sequences, wherein the observed or measured nucleotide incorporation events are obtained from a plurality of defined spaces of a chip or substrate in response to a nucleotide flow to the defined spaces, wherein each defined space is in communication with a respective sensor and wherein at least some of the defined spaces comprise the one or more sample nucleic acids, wherein each defined space is a space in which at least some of a molecule, fluid, and/or a solid can be confined, retained and/or localized, wherein the plurality of defined spaces comprises at least 100,000 defined spaces;
developing a plurality of predicted modelings of the observed or measured nucleotide incorporation events;
comparing the plurality of predicted modelings to at least one of the observed or measured nucleotide incorporation events; and
identifying at least a portion of the nucleic acid sequence on a basis of which predicted modeling is most similar to the observed or measured nucleotide incorporation events, the portion covering at least two consecutive bases called together as a whole.
18 . The method of claim 17 , wherein the comparing step applies a multi-element traversal to compare the predicted modelings to the observed or measured nucleotide incorporation events.
19 . The method of claim 18 , wherein the multi-element traversal comprises a tree-based search.
20 . The method of claim 18 , wherein the multi-element traversal includes one or more metrics to compare the predicted modelings to the observed or measured nucleotide incorporation events, the identifying step further comprising determining a path through the multi-element traversal having a best fit using the one or more metrics to determine which of the predicted modelings is most similar to the observed or measured nucleotide incorporation events.