Microbiome based systems, apparatus and methods for the exploration and production of hydrocarbons

1. A system comprising:

a database to store microbiome data corresponding to one or more sample materials obtained from a resource production field at one or more time points; and

a computer system configured to perform operations comprising:

analyzing the microbiome data to generate derived microbiome data corresponding to the one or more sample materials, the generating of the derived microbiome data including computationally linking the microbiome data to metadata corresponding to one or more industrial setting factors at the one or more time points;

selecting a predictive machine-learned model from a plurality of predictive machine-learned models based on an expected prediction error associated with the predictive machine-learned model, an accuracy of the expected prediction error for the predictive machine-learned model having been improved through a number of repeated tests of the expected predictive accuracy using multiple subsets of the derived microbiome data, the predictive machine-learned model having been generated using an embedded approach to feature selection;

generating predictive microbiome data corresponding to one or sample materials obtained at one or more additional time points, the generating including applying the selected predictive machine-learned model to derived microbiome data corresponding one or more sample materials obtained at the one or more additional time points; and

causing an interactive graphical user interface to be presented on a device, the interactive graphical user interface including an interactive tool for visualization of patterns in the predictive microbiome data corresponding to the one or more sample materials obtained at the one or more additional time points, the predictive microbiome data pertaining to at least one of production volume, fluid origin, fluid migration, oil saturation, water saturation, geological stratigraphy, temperature, viscosity, pressure, hydrocarbon composition, total organic carbon, permeability, or porosity.

2. The system of claim 1 , wherein the derived microbiome data includes diversity metrics and the generating of the predictive microbiome data includes determining a statistical significance of the diversity metrics to the regression.

3. The system of claim 1 , wherein the sample materials include at least one of liquids, soils, or rocks.

4. The system of claim 1 where the derived microbiome data includes environmental metadata pertaining to at least one of liquids, soils, or rocks collected from the field at the one or more points in time.

5. The system of claim 1 , wherein the derived microbiome data includes general metadata pertaining to at least one of physical characteristics, chemical characteristics, geological characteristics associated with the one or more sample materials.

6. The system of claim 1 , wherein the predictive microbiome data is utilized in conjunction with additional subsurface information comprising high resolution geologic maps and logging information, including at least one of gamma ray, triple combo, bit mechanics, derived lithology indicators, total organic content, or special core analyses information.

7. A method comprising:

accessing a database that stores microbiome data corresponding to one or more sample materials obtained from a resource production field at one or more time points;

analyzing the microbiome data to generate derived microbiome data corresponding to the one or more sample materials, the generating of the derived microbiome data including computationally linking the microbiome data to metadata corresponding to one or more industrial setting factors at the one or more time points;

selecting a predictive machine-learned model from a plurality of predictive machine-learned models based on an expected prediction error associated with the predictive machine-learned model, an accuracy of the expected prediction error for the predictive machine-learned model having been improved through a number of repeated tests of the expected predictive accuracy using multiple subsets of the derived microbiome data, the predictive machine-learned model having been generated using an embedded approach to feature selection;

generating predictive microbiome data corresponding to one or sample materials obtained at one or more additional time points, the generating including applying the selected predictive machine-learned model to derived microbiome data corresponding to the one or more sample materials obtained at the one or more additional time points; and

causing an interactive graphical user interface to be presented on a device, the interactive graphical user interface including an interactive tool for visualization of patterns in the predictive microbiome data corresponding to the one or more sample materials at the one or more additional time points, the predictive microbiome data pertaining to at least one of production volume, fluid origin, fluid migration, oil saturation, water saturation, geological stratigraphy, temperature, viscosity, pressure, hydrocarbon composition, total organic carbon, permeability, or porosity.

8. The method of claim 7 , wherein the derived microbiome data and the historical data includes diversity metrics and the generating of the predictive microbiome data includes determining a statistical significance of the comparison with respect to the diversity metrics.

9. The method of claim 7 , wherein the sample materials include at least one of liquids, soils, or rocks.

10. The method of claim 7 where the derived microbiome data includes environmental metadata pertaining to at least one of liquids, soils, or rocks collected from the field at the one or more points in time.

11. The method of claim 7 , wherein the derived microbiome data includes general metadata pertaining to at least one of physical characteristics, chemical characteristics, geological characteristics associated with the one or more sample materials.

12. The method of claim 8 , wherein the predictive microbiome data includes a prediction of a presence or absence of hydrocarbons at the one or more locations and the method further comprises communicating a visualization of the prediction of the presence or absence of hydrocarbons at the one or more locations along with the statistical significance to assist the operator in identification of possible patterns.

13. A non-transitory machine readable storage medium embodying instructions that, when executed by a processor, cause the processor to perform operations comprising:

accessing a database that stores microbiome data corresponding to one or more sample materials obtained from a resource production field at one or more time points;

analyzing the microbiome data to generate derived microbiome data corresponding to the one or more sample materials, the generating of the derived microbiome data including computationally linking the microbiome data to metadata corresponding to one or more industrial setting factors at the one or more time points;

selecting a predictive machine-learned model from a plurality of predictive machine-learned models based on an expected prediction error associated with the predictive machine-learned model, an accuracy of the expected prediction error for the predictive machine-learned model having been improved through a number of repeated tests of the expected predictive accuracy using multiple subsets of the derived microbiome data, the predictive machine-learned model having been generated using an embedded approach to feature selection;

generating predictive microbiome data corresponding to one or sample materials obtained at one or more additional time points, the generating including applying the selected predictive machine-learned model to derived microbiome data corresponding to the one or more sample materials obtained at the one or more additional time points; and

causing an interactive graphical user interface to be presented on a device, the interactive graphical user interface including an interactive tool for visualization of patterns in the predictive microbiome data corresponding to the one or more sample materials at the one or more additional time points, the predictive microbiome data pertaining to at least one of production volume, fluid origin, fluid migration, oil saturation, water saturation, geological stratigraphy, temperature, viscosity, pressure, hydrocarbon composition, total organic carbon, permeability, or porosity.

14. The non-transitory machine readable storage medium of claim 13 , wherein the derived microbiome data and the historical data includes diversity metrics and the generating of the predictive microbiome data includes determining a statistical significance of the comparison with respect to the diversity metrics.

15. The non-transitory machine readable storage medium of claim 13 , wherein the sample materials include at least one of liquids, soils, or rocks.

16. The non-transitory machine readable storage medium of claim 13 where the derived microbiome data includes environmental metadata pertaining to at least one of liquids, soils, or rocks collected from the field at the one or more points in time.

17. The non-transitory machine readable storage medium of claim 13 , wherein the derived microbiome data includes general metadata pertaining to at least one of physical characteristics, chemical characteristics, geological characteristics associated with the one or more sample materials.

18. The system of claim 1 , wherein the applying of the selected predictive model is based on an additional test of the expected prediction error using a test set was held out from the multiple subsets of the derived microbiome data.