Module-Compass

Module-Compass Quickstart

compass --data expression.tsv --model Human1 --species homo_sapiens --select-meta-subsystems meta_subsystems.txt --num-processes 10

COMPASS computes reaction scores for each reaction in the entire metabolic network, but in many cases users may only be interested in a given collection of subsystems (by which we refer to in this documentation as meta-subsystems) within the network.

The --select-meta-subsystems flag invokes Module-Compass and computes reaction scores on the meta-subsystems defined in a user-provided textual file. An example input file for Human1/Mouse1 is provided below:

CENTRAL_CARBON_META_SUBSYSTEM: Glycolysis / Gluconeogenesis; Tricarboxylic acid cycle and glyoxylate/dicarboxylate metabolism; Oxidative phosphorylation
AA_META_SUBSYSTEM: Phenylalanine, tyrosine and tryptophan biosynthesis; Arginine and proline metabolism; Tyrosine metabolism; Folate metabolism; Glycine, serine and threonine metabolism; Valine, leucine, and isoleucine metabolism; Cysteine and methionine metabolism; Lysine metabolism; Alanine, aspartate and glutamate metabolism; Metabolism of other amino acids; Histidine metabolism; Tryptophan metabolism; Beta-alanine metabolism; Phenylalanine metabolism; Urea cycle
LIPID_META_SUBSYSTEM: Fatty acid oxidation; Formation and hydrolysis of cholesterol esters; Arachidonic acid metabolism; Carnitine shuttle (endoplasmic reticular); Sphingolipid metabolism; Carnitine shuttle (mitochondrial); Glycerophospholipid metabolism; Steroid metabolism; Glycerolipid metabolism; Fatty acid biosynthesis; Omega-3 fatty acid metabolism; Acyl-CoA hydrolysis; Fatty acid activation (cytosolic); Cholesterol metabolism; Omega-6 fatty acid metabolism; Carnitine shuttle (cytosolic); Beta oxidation of even-chain fatty acids (mitochondrial); Fatty acid activation (endoplasmic reticular); Fatty acid metabolism; Glycosphingolipid biosynthesis-lacto and neolacto series; Beta oxidation of even-chain fatty acids (peroxisomal); Beta oxidation of unsaturated fatty acids (n-9) (peroxisomal); Beta oxidation of odd-chain fatty acids (mitochondrial); Glycosphingolipid biosynthesis-ganglio series; Linoleate metabolism; Fatty acid biosynthesis (even-chain); Fatty acid biosynthesis (unsaturated); Fatty acid biosynthesis (odd-chain); Cholesterol biosynthesis 1 (Bloch pathway); Glycosphingolipid metabolism; Propanoate metabolism; Beta oxidation of poly-unsaturated fatty acids (mitochondrial); Beta oxidation of unsaturated fatty acids (n-9) (mitochondrial); Phosphatidylinositol phosphate metabolism; Fatty acid elongation (even-chain); Beta oxidation of phytanic acid (peroxisomal); Beta oxidation of di-unsaturated fatty acids (n-6) (peroxisomal); Beta oxidation of di-unsaturated fatty acids (n-6) (mitochondrial); Beta oxidation of unsaturated fatty acids (n-7) (mitochondrial); Ether lipid metabolism; Cholesterol biosynthesis 3 (Kandustch-Russell pathway); Beta oxidation of branched-chain fatty acids (mitochondrial); Fatty acid elongation (odd-chain); Fatty acid desaturation (even-chain); Glycosphingolipid biosynthesis-globo series; Ubiquinone synthesis; Butanoate metabolism; Carnitine shuttle (peroxisomal); Beta oxidation of unsaturated fatty acids (n-7) (peroxisomal); Ubiquinone and other terpenoid-quinone biosynthesis; Lipoic acid metabolism; Cholesterol biosynthesis 2; Fatty acid desaturation (odd-chain); Beta oxidation of odd-chain fatty acids (peroxisomal); Triacylglycerol synthesis; Fatty acid degradation

An example input file for RECON2 is provided below:

CENTRAL_CARBON_META_SUBSYSTEM: Glycolysis/gluconeogenesis; Citric acid cycle; Oxidative phosphorylation
AA_META_SUBSYSTEM: Alanine and aspartate metabolism; Arginine and Proline Metabolism; beta-Alanine metabolism; Cysteine Metabolism; D-alanine metabolism; Folate metabolism; Glutamate metabolism; Glycine, serine, alanine and threonine metabolism; Histidine metabolism; Lysine metabolism; Methionine and cysteine metabolism; Taurine and hypotaurine metabolism; Tryptophan metabolism; Tyrosine metabolism; Urea cycle; Valine, leucine, and isoleucine metabolism
LIPID_META_SUBSYSTEM: Fatty acid oxidation; Glycerophospholipid metabolism; Fatty acid synthesis; Steroid metabolism; Sphingolipid metabolism; Triacylglycerol synthesis; Cholesterol metabolism; Propanoate metabolism; Squalene and cholesterol synthesis; Ubiquinone synthesis; Phosphatidylinositol phosphate metabolism; Butanoate metabolism

Each line in the file defines a meta-subsystem which should be formatted as:

[META_SUBSYSTEM_ID]: [SUBSYSTEM_1]; [SUBSYSTEM_2]; ...

Note that the meta-subsystem IDs should not contain spaces. Also note that the subsystems contained in each meta-subsystem should be separated by semicolons.

If you are using Human1, a list of subsystems supported by Module-Compass can be found here. A list of reactions supported by Module-Compass can be found here.

If you are using Mouse1, a list of subsystems supported by Module-Compass can be found here. A list of reactions supported by Module-Compass can be found here.

If you are using RECON2, a list of subsystems supported by Module-Compass can be found here. A list of reactions supported by Module-Compass can be found here.

Module-Compass Output

The output of Module-Compass should be similar to the following:

|-- <output-dir>
|   |-- <temp-dir>
|   |   |-- <META_SUBSYSTEM_ID_1>
|   |   |-- <META_SUBSYSTEM_ID_2>
|   |-- <META_SUBSYSTEM_ID_1>
|   |-- <META_SUBSYSTEM_ID_2>
|   |-- compass.log
|   |-- meta_subsystem_cache
|   |   |-- <META_SUBSYSTEM_ID_1>
|   |   |-- <META_SUBSYSTEM_ID_2>
|   |-- meta_subsystem_models
|   |   |-- <META_SUBSYSTEM_ID_1>
|   |   |-- <META_SUBSYSTEM_ID_2>

<temp-dir> contains the intermediate results for each meta-subsystem, similar to a regular Compass run.

<META_SUBSYSTEM_ID> folders contain the reactions.tsv file that contains reactions scores for reactions in each meta-subsystem.

meta_subsystem_cache contains the optimal reaction flux for each reaction. Optimal fluxes are computed agnostic of gene expression and are only based on the connectivity of reactions within the metabolic network. Please refer to the manuscript for a detailed treatment of the inner workings of Compass.

meta_subsystem_models contains the reactions directly associated with the meta-subsystem, their one-hop neighbors, and relevant exchange reactions. The format may differ between various choices of metabolic model (e.g. RECON2 or Human1), but should be similar to the corresponding comprehensive model under the Metabolic Models directory.

An example reactions.tsv output file is shown below:

For each meta-subsystem, Module-Compass computes reaction scores for two types of reactions: those that are directly associated with the meta-subsystem, as well as exchange reactions for metabolites associated with these reactions. The former are labeled based on their reaction IDs in the metabolic model, while the latter are formatted as <MET_ID>_EXCHANGE_<META_SUBSYSTEM_ID>. The metadata for these reactions can be found under the corresponding folder for the meta-subsystem of interest within the meta_subsystem_models directory.

Module-Compass Explained

To support running COMPASS on a subset of the network, we hereby provide Module-Compass, an algorithm that treats user-defined meta-subsystems as individual networks but still provides the necessary context to ensure that this simplification of the network is reasonable.

More specifically, each meta-subsystem is defined as a collection of subsystems that the user chooses. We construct the network for each meta-subsystem by first taking all reactions associated with the subsystems within the given meta-subsystems, then taking their one-step neighbor reactions, and finally adding exchange reactions for all metabolites associated with the collection of reactions we constructed above. The neighboring reactions provides necessary context to the meta-subsystem to ensure a reasonable output, while the exchange reactions enables us to treat the constructed network as one that is able to interface with the surrounding environment and uptake/secrete metabolites.

One thing to note is that the --select-reactions and --select-subsystems flags also allow users to specify a list of reactions and subsystems to compute COMPASS on. However, Module-Compass differs fundamentally from these methods. --select-reactions and --select-subsystems both operate on the entire metabolic network, meaning that the computation of each reaction score is still constrained by all reactions in the entire metabolic network. This requires the linear solver to optimize for all variables (reactions). On the other hand, Module-Compass treats meta-subsystems as standalone networks, drastically reducing the number of variables in the linear optimization problem and therefore resulting in significant speedups.