Day 2: Environmental Metabolomics¶

This tutorial narrative demonstrates how to¶

• Apply thermodynamic theory (known as λ theory) to convert molecular formulas of compounds (derived from FTICR-MS peaks using Formulatiry and R codes) into stoichiometric and kinetic forms of biogeochemical reactions
• Use the resulting kinetic equations to simulate dynamic conversion of compounds in batch and continuous stirred tank reactors

Main questions to address¶

• Under what conditions, are respiration rates driven by thermodynamics?
• How do respiration rates respond to the variations in parameters and input variables and how to interpret the results?

Step-by-step implementation with specific questions¶

Note that most of the apps used in this narrative are beta versions. For apps that cannot be found from 'R' (release) versions, click 'R' to switch over to 'B' (beta) versions on the App category.

Part I: Characterize theromodynamic properties of compounds through λ distribution

Step 1: Upload FTICR data files

• Step 1a: Add WHONDRS FTICR data files (generated from Formularity and R-code) into the staging area (drag and drop data files)
• Step 1b: Import the data from the staging area into the narrative (App: Import TSV/Excel File as Attribute Mapping from Staging)

Step 2: Generate thermodynamic properties and stoichiometric reactions from molecular formulas of compounds (App: Build Stoichiometric Reaction Models From Chemical Formulas)

• How do you define ΔG_Cox⁰ and λ?
• How can the mean and median values and standard deviation of lambda distribution characterize thermodynamic properties of compounds?
• Can the value of lambda be zero? If so, what does that mean?
• What is the output object 'stoichMet_O2_xxx'?
• What is the output object 'Bin_Averaged_stoichMet_O2_xxx'?
• How are average molecular formulas in each lambda bin calculated?

Part II: Understand the relationships between λ and reaction rates

Step 3: Upload lambda and stoichiometric reaction files

• Step 3a: Download ‘thermodynamic_props.csv’ and ‘stoichMet_O2.csv’ from the output files of the previous step
• Step 3b: Add these two files into the staging area (drag and drop);
• Step 3c: Import data from the staging area into the narrative (App: Import TSV/Excel File as Attribute Mapping from Staging)
• What is the result file 'thermodynamic_props.csv'?
• What is the result file 'stoichMet_O2.csv'?

Step 4: Examine the correlations between λ and reaction rates (App: Run Lambda Analysis)

• What input variables/parameters are required to plot reaction rate distributions? Why does the app require two input variables (v_h[Cs] and v_h[O₂]), instead of three (v_h, [Cs], and [O₂])?
• Do the correlations of lambda with reaction rates vary among r_O2, r_Biom, and If so (or not), what does that mean?
• Under what conditions, lambda has good or poor correlations with reaction rates? What do these results mean?

Part III: Compare dynamic simulations in batch and continuous reactors and analyze how simulation outputs respond to parameter variations¶

Step 5: Simulate biogeochemical models in a batch reactor (App: Simulate Batch Biogeochemical Reaction Model)

• What are the units of input parameters and variables: dilution rate (D), μ^max, V_h, k_d, [Cs_feed] (and [Cs₀]), [O_2,feed] (and [O_2,in] and what are their meanings?
• How do you define “specific” reaction rates? What are their units?
• How do you compare the results between randomly selected compound set vs. lambda bin-based average chemical formulas?
• What is the role of cybernetic regulation in the model? How and why are the values of specific reaction rates different, with vs. without turning on the cybernetic regulation option?
• How do simulation outputs vary in response to the changes in parameters and input variables? How do you interpret them?

Step 6: Simulate biogeochemical models in a continuous stirred tank reactor (App: Simulate CSTR Biogeochemical Reaction Model)

• What input parameters and variables are additionally required in the CSTR model, compared to the batch model?
• What inputs are required for batch simulations but NOT required for CSTR simulations?
• Why are compound conversions in CSTR lower than batch? What parameters can be changed to increase/decrease yields?
• Answer the last three questions in Step 5.

Step 7: Repeat Steps 5 and 6 with λ bin-averaged compounds

Questions? Contact Hyun-Seob Song (hsong5@unl.edu) or Joon-Yong Lee (joonyong.lee@pnnl.gov)!