Supplementary information to

One-shot ¹³C¹⁵N-metabolic flux analysis for simultaneous quantification of carbon and nitrogen flux

Khushboo Borah Slater^1,¶, Martin Beyß^2,3,¶, Ye Xu¹, Jim Barber¹, Catia Costa⁴, Jane Newcombe¹, Axel Theorell^2,5, Melanie J Bailey⁴, Dany JV Beste¹, Johnjoe McFadden^1,¶,*, Katharina Nöh^2,¶,*

¹ Faculty of Health and Medical Sciences, University of Surrey, UK, GU2 7XH

² Forschungszentrum Jülich GmbH, Institute of Bio- and Geosciences, IBG-1: Biotechnology, 52425 Jülich, Germany

³ RWTH Aachen University, Computational Systems Biotechnology, 52062 Aachen, Germany

⁴ Faculty of Engineering and Physical Sciences, University of Surrey, UK, GU2 7XH

⁵ Current address: Computational Systems Biology, ETH Zürich, 4058 Basel, Switzerland

^¶ These authors contributed equally to this work

^* Correspondence email address: [email protected]; [email protected]

Contents

• x3cflux-bma-1.0.0-Ubuntu2204.deb Bayesian Model Averaging extension to the high-performance simulator 13CFLUX2. Precondition of use: a functional installation of 13CFLUX2 (Ubuntu 22.04.).
• CN_network_model_BCG.fml FluxML model of BCG used for the analysis including reactions, CN atom transitions, constraints, and data
• run.sh shell script to execute Bayesian Model Averaging in parallel (50 parallel processes) with parameters described in the publication
• seeds.txt the seeds used for producing the results
• run_seeds.sh shell script to execute Bayesian Model Averaging as described before, but uses the file seeds.txt for the seeds. Executing the script Will reproduce the samples analyzed in the publication.
• README.md this document. Guide to x3cflux-bma.

Usage

To reproduce the samples from the publication the following is needed:

• 13CFLUX2 version 2.3 or newer. 13CFLUX2 is available free of charge for non-commercial academic and research purposes. Further details are available at the 13CFLUX2 website.
• Bayesian Model Averaging extension to 13CFLUX2, which is available in this repository for Ubuntu 22.04.

Clone the repository and execute either run_seeds.sh (to use the same seeds as used in the publication) or run.sh (to use new seeds).

Parameters to x3cflux-bma

CLI parameter	description	value used in publication
--thinning <int>	Save only each thinning datapoint to decrease size on disc of saved data (default: 1)	2000
--fullInfo <bool>	Save independent and dependent states, simulated measurements and likelihoods (default: false)	1
-o <HDF5 file>, --output <HDF5 file>	(required) File name of HDF5 output file	autogenerated from script
--start-fml	Starts all chains from the flux distribution of the FML file (default: disabled)	-
-j <int>, --jobs <int>	Number of jobs (default: number of virtual cores)	12 (does impact runtime, but not results)
--seed <int>	Random number generator seed (default: current time)	- (the seeds uses are available in seeds.txt)
--chains <int>	Number of chains (default: 20)	30
-p <double>, --pw <double>	Width of proposal Gaussian in RW (default: 2.5)	50
-b <int>, --burnin <int>	Number of burnin samples (default: 20000)	5,000,000
-n <int>, --samples <int>	Number of samples, excluding burn in (default: 10000)	10,000,000
-i <FML>, --input <FML>	(required) FML file	autogenerated from script

interpreting results

x3cflux-bma produces hdf5 files. Relevant for the evaluation are the data in the samples branch

• likelihoods likelihood (SSR/2) of sample
• states the samples, i.e. values of the free fluxes; exchange fluxes are scaled to [0, 1]
• coIs names of the free fluxes
• netFluxes values of all net fluxes
• allFluxCoIs names of all net fluxes