System and method for analysis of biological data
Inventors
Oved, Kfir • Eden, Eran • NAVON, Roy • COHEN-DOTAN, Assaf • KRONENFELD, Gali • BOICO, Olga
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-12205677-B2
Publication Date
2025-01-21
Expiration Date
Abstract
A method of analyzing biological data is disclosed. The method comprises obtaining biological data containing at least an expression level of MX dynamin-like GTPase 1 (MX1) and an expression level of C-reactive protein (CRP) in the blood of a subject, calculating a distance between a segment of a curved line and an axis defined by a direction, the distance being calculated at a point over the curved line defined by a coordinate along the direction, and correlating the distance to the presence of, absence of, or likelihood that the subject has, a bacterial infection.
Core Innovation
The invention analyzes biological data by obtaining blood expression levels of MX1 and C-reactive protein (CRP) in a subject. A hardware processor calculates a distance between a segment of a curved line and an axis defined by a direction, where the distance is calculated perpendicularly to the axis and a coordinate δ along the direction defines the curved line.
In one embodiment, δ equals a0+a1X+a2Z or a0+a1X+a2Y, where X is a value of CRP in μg/ml and ZMX1 is a z-score of MX1 relative to a group of subjects previously diagnosed with a bacterial infection, or Y is a value of MX1 measured by flow cytometry. The method stores the calculated distance and correlates the distance to the presence of, absence of, or likelihood that the subject has a bacterial infection.
The method generates an output on a graphical user interface of the presence, absence, or likelihood, compares the likelihood to a predetermined threshold, and treats the subject for the bacterial infection with an antibiotic agent when the likelihood is above the predetermined threshold. The geometric computation is constrained by requiring that at least 90% of the segment is between a lower bound line f(δ)−ε0 and an upper bound line f(δ)+ε1, where f(δ) equals 1/(1+exp(−δ)) and the parameters a0, a1, and a2 are limited to specified ranges.
Claims Coverage
The document provides three independent methods that each analyze biological data using MX1 and CRP in blood, compute a perpendicular distance from a curved line segment to an axis via a coordinate δ, correlate the distance to bacterial infection presence, absence, or likelihood, output the result, compare to a predetermined threshold, and treat with an antibiotic when the likelihood exceeds the threshold. Each independent method further constrains the curved-line segment coverage within logistic-function bounds and specifies the relationship of δ to CRP and MX1, including different definitions for the MX1 component and/or the measurement modality.
Curved-line distance from CRP and MX1 z-scored coordinate
calculating by a hardware processor a distance between a segment of a curved line and an axis defined by a direction, said distance being calculated perpendicularly to said axis, between a point on said axis and a corresponding point over said curved line, said points being defined by a coordinate δ along said direction, wherein said coordinate δ, once calculated, equals a0+a1X+a2Z, and wherein said X is a value of said CRP in μg/ml, and said ZMX1 is a z-score of said MX1 relative to a group of subjects previously diagnosed with a bacterial infection
Distance-to-bacterial-infection likelihood correlation with threshold antibiotic treatment
by said hardware processor, correlating said distance to the presence of, absence of, or likelihood that the subject has, a bacterial infection; generating on a graphical user interface an output of said presence, absence or likelihood; obtaining said likelihood based on said distance; comparing said likelihood to a predetermined threshold; and treating the subject for said bacterial infection with an antibiotic agent when said likelihood is above said predetermined threshold
Logistic-bound segment constraint for the curved-line distance
wherein at least 90% of said segment is between a lower bound line f(δ)−ε0 and an upper bound line f(δ)+ε1, wherein said f(δ) equals 1/(1+exp(−δ)), wherein each of said ε0 and said ε1 is less than 0.5, and wherein a0 is from about −2.4 to about −1.9, a1 is from about 0.04 to about 0.05, and a2 is from about −0.39 to about −0.43
Curved-line distance using CRP and MX1 values in linear coordinate δ
calculating by a hardware processor a distance between a segment of a curved line and an axis defined by a direction, said distance being calculated perpendicularly to said axis, between a point on said axis and a corresponding point over said curved line, said points being defined by a coordinate δ along said direction, wherein said coordinate δ, once calculated, equals a0+a1X+a2Y, and wherein said X is a value of said CRP in μg/ml, and said Y is a value of said MX1 in ng/ml
Distance-to-bacterial-infection likelihood correlation with threshold antibiotic treatment
by said hardware processor, correlating said distance to the presence of, absence of, or likelihood that the subject has, a bacterial infection; generating on a graphical user interface an output of said presence, absence or likelihood; obtaining said likelihood based on said distance; comparing said likelihood to a predetermined threshold; and treating the subject for said bacterial infection with an antibiotic agent when said likelihood is above said predetermined threshold
Logistic-bound segment constraint with specified δ-parameter ranges
wherein at least 90% of said segment is between a lower bound line f(δ)−ε0 and an upper bound line f(δ)+ε1, wherein said f(δ) equals 1/(1+exp(−δ)), wherein each of said ε0 and said ε1 is less than 0.5, and wherein a0 is from about 0.4 to about 0.5, a1 is from about 0.015 to about 0.02, and a2 is from about −0.0025 to about −0.0018
Curved-line distance using flow cytometry MX1 value in linear coordinate δ
calculating by a hardware processor a distance between a segment of a curved line and an axis defined by a direction, said distance being calculated perpendicularly to said axis, between a point on said axis and a corresponding point over said curved line, said points being defined by a coordinate δ along said direction, wherein said coordinate δ, once calculated, equals a0+a1X+a2Y, and wherein said X is a value of said CRP in μg/ml, and said Y is a value of said MX1 when measured by flow cytometry
Distance-to-bacterial-infection likelihood correlation with threshold antibiotic treatment
by said hardware processor, correlating said distance to the presence of, absence of, or likelihood that the subject has, a bacterial infection; generating on a graphical user interface an output of said presence, absence or likelihood; obtaining said likelihood based on said distance; comparing said likelihood to a predetermined threshold; and treating the subject for said bacterial infection with an antibiotic agent when said likelihood is above said predetermined threshold
Logistic-bound segment constraint with specified δ-parameter ranges
wherein at least 90% of said segment is between a lower bound line f(δ)−ε0 and an upper bound line f(δ)+ε1, wherein said f(δ) equals 1/(1+exp(−δ)), wherein each of said ε0 and said ε1 is less than 0.5, and wherein a0 is from about −1.7 to about −1.4, a1 is from about 0.03 to about 0.05, and a2 is from about −5.8E−205 to about −4.7E−205
The independent claims collectively cover a computer-implemented diagnostic workflow that turns blood expression levels of MX1 and CRP into a distance metric derived from a curved line and an axis using a coordinate δ defined by stated coefficients. The computed distance is correlated to bacterial-infection presence/absence or likelihood, displayed, thresholded, and used to determine antibiotic treatment when the likelihood exceeds the predetermined threshold, subject to explicit geometric and logistic bounds.
Stated Advantages
Provides an output of the presence, absence, or likelihood of bacterial infection on a graphical user interface.
Enables clinical decision-making by comparing the likelihood to a predetermined threshold and treating with an antibiotic agent when the likelihood is above the threshold.
Supports diagnostic performance for separating bacterial infection categories as described in the clinical study example, including AUC, sensitivity, and specificity values for bacterial versus viral and infectious versus non-infectious comparisons.
Supports sepsis ruling in using the MX1+CRP distance in combination with respiratory rate, mental state, and blood pressure/SOFA criteria.
improving sensitivity important for antibiotic guidance
Documented Applications
Analyzing blood expression levels of MX1 and CRP to determine the presence, absence, or likelihood of bacterial infection for use in clinical decision-making with a predetermined threshold and antibiotic treatment.
Sepsis use case described as ruling in sepsis using MX1+CRP distance in combination with respiratory rate, mental state, and blood pressure/SOFA criteria.
Validation across clinical syndromes, pathogens, and comorbidities, including analysis of infectious versus non-infectious exacerbation state in COPD.
Distinguishing acute bacterial vs viral infections using host-response proteins (including MX1 and CRP) in a multicenter prospective clinical study.
Diagnostic discrimination for bacterial infection vs non-bacterial conditions (including infectious vs non-infectious) using logistic-regression models that combine CRP and MX1.
Robustness evaluation across MX1 measurement techniques (flow cytometry vs sandwich-type immunoassay) and performance reporting by clinical syndromes/pathogen.
Using cutoff-independent logistic models that outperform CRP cutoff-dependent schemes to improve sensitivity for antibiotic guidance.
Interested in licensing this patent?