Abstract: One or more techniques for optimizing cancer classification based on covariate characteristics is disclosed. In a first approach, an analytics system may determine separate cutoff thresholds for positively detecting disease signal for different labels for a covariate characteristic. The system may subdivide training samples based on their labels for the covariate characteristic, to separately determine the cutoff thresholds. In other approaches, the system may train disparate classifiers for each population. The system separates the training samples based on their labels for the covariate characteristic, and separately trains classifiers to generate a signal vector representing an amount of disease signal detected in a sample. The classifiers may be trained on different feature sets as determined based on mutual information gain, genomic region coverage, and healthy activation fraction.

Abstract: Systems and methods for determining a cancer class of a subject are provided in which a plurality of sequence reads, in electronic form, are obtained from a biological sample of the subject. The sample comprises a plurality of cell-free DNA molecules including respective DNA molecules longer than a threshold length of less than 160 nucleotides. The plurality of sequence reads excludes sequence reads of cell-free DNA molecules in the plurality of cell-free DNA molecules longer than the threshold length. The plurality of sequence reads is used to identify a relative copy number at each respective genomic location in a plurality of genomic locations in the genome of the subject. The genetic information about the subject obtained from the sample and the genetic information consisting of the identification of the relative copy number at each respective genomic location, is applied to a classifier that determines the cancer class of the subject.

Abstract: Classification of cancer condition, in a plurality of different cancer conditions, for a species, is provided in which, for each training subject in a plurality of training subjects, there is obtained a cancer condition and a genotypic data construct including genotypic information for the respective training subject. Genotypic constructs are formatted into corresponding vector sets comprising one or more vectors. Vector sets are provided to a network architecture including a convolutional neural network path comprising at least a first convolutional layer associated with a first filter that comprise a first set of filter weights and a scorer. Scores, corresponding to the input of vector sets into the network architecture, are obtained from the scorer. Comparison of respective scores to the corresponding cancer condition of the corresponding training subjects is used to adjust the filter weights thereby training the network architecture to classify cancer condition.

Abstract: Systems and methods for detecting a subtype of a disease state and for determining the development of a resistance mechanism in a disease are disclosed. One method may include: receiving, at an input component of the system, a set of sequence reads associated with a nucleic acid sample; generating, using a processor of the system and via analysis of the set of sequence reads, methylation data; and analyzing, using the processor, the methylation data to identify the subtype of the disease state. Another method may include: obtaining methylation data from a targeted methylation sequencing assay, applying the methylation data to a trained machine learning model, and receiving an output indicating whether MRD is present in a test subject and/or whether a resistance mechanism has been developed by a disease. Other aspects are described and claimed.

Abstract: Systems and methods for reducing noise for the analysis of low coverage sequencing data from a nucleic acid sample using a method, including: receiving, at an input component of the system, a set of sequence reads associated with the nucleic acid sample; allocating, using a processor component of the system, the set of sequence reads into a plurality of genomic bins; and introducing, subsequent to the allocating, a pseudocount number to bincount values to produce a smoothed dataset, wherein each of the bincount values is associated with one of the plurality of genomic bins. Other aspects are described and claimed.

Abstract: Classification of cancer condition, in a plurality of different cancer conditions, for a species, is provided in which, for each training subject in a plurality of training subjects, there is obtained a cancer condition and a genotypic data construct including genotypic information for the respective training subject. Genotypic constructs are formatted into corresponding vector sets comprising one or more vectors. Vector sets are provided to a network architecture including a convolutional neural network path comprising at least a first convolutional layer associated with a first filter that comprise a first set of filter weights and a scorer. Scores, corresponding to the input of vector sets into the network architecture, are obtained from the scorer. Comparison of respective scores to the corresponding cancer condition of the corresponding training subjects is used to adjust the filter weights thereby training the network architecture to classify cancer condition.

Abstract: Technical solutions for classifying patients with respect to multiple cancer classes are provided. The classification can be done using cell-free whole genome sequencing information from subjects. A reference set of subjects is used to train classifiers to recognize genomic markers that distinguish such cancer classes. The classifier training includes dividing the reference genome into a set of non-overlapping bins, applying a dimensionality reduction method to obtain a feature set, and using the feature set to train classifiers. For subjects with unknown cancer class, the trained classifiers provide probabilities or likelihoods that the subject has a respective cancer class for each cancer in a set of cancer classes. The present disclosure thus describes methods to improve the screening and detection of cancer class from among several cancer classes. This serves to facilitate early and appropriate treatment for subjects afflicted with cancer.

Abstract: Classification of cancer condition, in a plurality of different cancer conditions, for a species, is provided in which, for each training subject in a plurality of training subjects, there is obtained a cancer condition and a genotypic data construct including genotypic information for the respective training subject. Genotypic constructs are formatted into corresponding vector sets comprising one or more vectors. Vector sets are provided to a network architecture including a convolutional neural network path comprising at least a first convolutional layer associated with a first filter that comprise a first set of filter weights and a scorer. Scores, corresponding to the input of vector sets into the network architecture, are obtained from the scorer. Comparison of respective scores to the corresponding cancer condition of the corresponding training subjects is used to adjust the filter weights thereby training the network architecture to classify cancer condition.

Abstract: Technical solutions for classifying patients with respect to multiple cancer classes are provided. The classification can be done using cell-free whole genome sequencing information from subjects. A reference set of subjects is used to train classifiers to recognize genomic markers that distinguish such cancer classes. The classifier training includes dividing the reference genome into a set of non-overlapping bins, applying a dimensionality reduction method to obtain a feature set, and using the feature set to train classifiers. For subjects with unknown cancer class, the trained classifiers provide probabilities or likelihoods that the subject has a respective cancer class for each cancer in a set of cancer classes. The present disclosure thus describes methods to improve the screening and detection of cancer class from among several cancer classes. This serves to facilitate early and appropriate treatment for subjects afflicted with cancer.

Abstract: Classification of cancer condition, in a plurality of different cancer conditions, for a species, is provided in which, for each training subject in a plurality of training subjects, there is obtained a cancer condition and a genotypic data construct including genotypic information for the respective training subject. Genotypic constructs are formatted into corresponding vector sets comprising one or more vectors. Vector sets are provided to a network architecture including a convolutional neural network path comprising at least a first convolutional layer associated with a first filter that comprise a first set of filter weights and a scorer. Scores, corresponding to the input of vector sets into the network architecture, are obtained from the scorer. Comparison of respective scores to the corresponding cancer condition of the corresponding training subjects is used to adjust the filter weights thereby training the network architecture to classify cancer condition.

Abstract: Disclosed herein are systems and methods for localization of a disease state (e.g., tissue of origin of cancer) using nucleic acid samples. In an embodiment, a method comprises receiving a plurality of cancer signals of a sample, each cancer signal indicating a probability that the sample is associated with a different disease state of a plurality of disease states. The method determines a first cancer signal having a greatest probability among the plurality of cancer signals. In accordance with a determination that the first cancer signal satisfies a criterion, the method associates the sample with a first disease state. In accordance with a determination that the first cancer signal does not satisfy the criterion, the method determines a second cancer signal having a second greatest probability among the plurality of cancer signals, and associates the sample with the first disease state and a second disease state.

Abstract: Detecting cross-contamination between test samples used for determining cancer in a subject is beneficial. To detect cross-contamination, test sequences including at least one single nucleotide polymorphism are prepared using genome sequencing techniques. Some of the test sequences can be filtered to improve accuracy and precision. A prior contamination probability for each test sequence is determined based on a minor allele frequency. A contamination model including a likelihood test is applied to a test sequence. The likelihood test obtains a current contamination probability representing the likelihood that the test sample is contaminated. The contamination model can also determine a likelihood that the sample includes loss of heterozygosity representing the likelihood that the test sequence is contaminated. Test samples that are contaminated are removed. A source for the contaminated test sample can be found by comparing contaminated test sequences to other test sequences.

Abstract: Systems and methods described herein include detecting a presence or absence of HPV in a biological sample having cell-free nucleic acids from a subject and potentially cell-free nucleic acids from an HPV strain. Based on a detection of HPV viral nucleic acids in the biological sample, an HPV-based multiclass classifier that predicts a score for each HPV-associated cancer type is applied. The HPV-based multiclass classifier is trained on a training set of HPV-positive cancer samples. An HPV-associated cancer associated with the biological sample is determined based on the scores predicted by the HPV multiclass classifier.

Abstract: A method of identifying a plurality of features for estimating subject cell source fraction is provided. For each respective training subject in a plurality of training subjects, a corresponding methylation pattern of each respective cell-free fragment in a corresponding training plurality of cell-free fragments and a corresponding subject cancer indication is obtained. Each cell-free fragment is mapped to a bin in a plurality of bins, each bin representing a portion of a human reference genome. A cell-free fragment cancer condition is assigned to each cell-free fragment, as a function of a classifier upon inputting a corresponding methylation pattern of the respective cell-free fragment into the classifier. A measure of association is determined for each bin between the subject cancer condition and the cell-free fragment cancer condition. The plurality of features for estimating subject cell source fraction are identified as a subset of the plurality of bins.

Abstract: Methods and systems for detecting cancer and/or determining a cancer tissue of origin are disclosed. In some embodiments, a multiclass cancer classifier is disclosed that is trained with a plurality of biological samples containing cfDNA fragments. The analytics system derives a feature vector for each sample, and the multiclass classifier predicts a probability likelihood for each of a plurality of tissue of origin (TOO) classes. In some embodiments, the plurality of TOO classes include hematological subtypes, including both hematological malignancies and precursor conditions. In one embodiment, non-cancer samples having high tissue signal are pruned from the training sample set. In another embodiment, the analytics system stratifies samples according to tissue signal and applies binary threshold cutoffs determined for each stratum.

Abstract: Systems and methods for determining a cancer class of a subject are provided in which a plurality of sequence reads, in electronic form, are obtained from a biological sample of the subject. The sample comprises a plurality of cell-free DNA molecules including respective DNA molecules longer than a threshold length of less than 160 nucleotides. The plurality of sequence reads excludes sequence reads of cell-free DNA molecules in the plurality of cell-free DNA molecules longer than the threshold length. The plurality of sequence reads is used to identify a relative copy number at each respective genomic location in a plurality of genomic locations in the genome of the subject. The genetic information about the subject obtained from the sample and the genetic information consisting of the identification of the relative copy number at each respective genomic location, is applied to a classifier that determines the cancer class of the subject.

Abstract: Technical solutions for classifying patients with respect to multiple cancer classes are provided. The classification can be done using cell-free whole genome sequencing information from subjects. A reference set of subjects is used to train classifiers to recognize genomic markers that distinguish such cancer classes. The classifier training includes dividing the reference genome into a set of non-overlapping bins, applying a dimensionality reduction method to obtain a feature set, and using the feature set to train classifiers. For subjects with unknown cancer class, the trained classifiers provide probabilities or likelihoods that the subject has a respective cancer class for each cancer in a set of cancer classes. The present disclosure thus describes methods to improve the screening and detection of cancer class from among several cancer classes. This serves to facilitate early and appropriate treatment for subjects afflicted with cancer.

Abstract: Classification of cancer condition, in a plurality of different cancer conditions, for a species, is provided in which, for each training subject in a plurality of training subjects, there is obtained a cancer condition and a genotypic data construct including genotypic information for the respective training subject. Genotypic constructs are formatted into corresponding vector sets comprising one or more vectors. Vector sets are provided to a network architecture including a convolutional neural network path comprising at least a first convolutional layer associated with a first filter that comprise a first set of filter weights and a scorer. Scores, corresponding to the input of vector sets into the network architecture, are obtained from the scorer. Comparison of respective scores to the corresponding cancer condition of the corresponding training subjects is used to adjust the filter weights thereby training the network architecture to classify cancer condition.

Abstract: A system, method and computer program product for analyzing data of high dimensionality (e.g., sequence reads of nucleic acid samples in connection with a disease condition) are provided.