Microbial 'omics

Brought to you by

# kegg-metabolism [artifact]

A TXT-type anvi’o artifact. This artifact is typically generated, used, and/or exported by anvi’o (and not provided by the user)..

🔙 To the main page of anvi’o programs and artifacts.

## Provided by

anvi-estimate-metabolism

## Required or used by

anvi-compute-metabolic-enrichment

## Description

Output text files produced by anvi-estimate-metabolism that describe the presence of metabolic pathways in a contigs-db.

Depending on the output options used when running anvi-estimate-metabolism, these files will have different formats. This page describes and provides examples of the various output file types.

## Long-format output modes

The long-format output option produces tab-delimited files. Each line in the file (except for the header) is indexed by an integer in the unique_id column. Different output “modes” will result in output files with different information.

### ‘Modules’ Mode

The modules mode output file will have the suffix modules.txt. Each line in the file will represent information about a KEGG module in a given genome, bin, or contig of a metagenome assembly. Here is one example, produced by running metabolism estimation on the Infant Gut dataset:

unique_id db_name genome_name kegg_module module_name module_class module_category module_subcategory module_definition module_completeness module_is_complete kofam_hits_in_module gene_caller_ids_in_module warnings
0 E_faecalis_6240 Enterococcus_faecalis_6240 M00001 Glycolysis (Embden-Meyerhof pathway), glucose => pyruvate Pathway modules Carbohydrate metabolism Central carbohydrate metabolism ””“(K00844,K12407,K00845,K00886,K08074,K00918) (K01810,K06859,K13810,K15916) (K00850,K16370,K21071,K00918) (K01623,K01624,K11645,K16305,K16306) K01803 ((K00134,K00150) K00927,K11389) (K01834,K15633,K15634,K15635) K01689 (K00873,K12406)””” 1.0 True K01834,K00134,K00873,K01689,K01624,K01803,K00927,K00850,K01810,K00845 2342,2646,1044,642,226,1041,348,1042,1043,225,600,1608 None
1 E_faecalis_6240 Enterococcus_faecalis_6240 M00002 Glycolysis, core module involving three-carbon compounds Pathway modules Carbohydrate metabolism Central carbohydrate metabolism ”"”K01803 ((K00134,K00150) K00927,K11389) (K01834,K15633,K15634,K15635) K01689 (K00873,K12406)””” 1.0 True K01834,K00134,K00873,K01689,K01803,K00927 2342,2646,1044,642,226,1041,1042,1043 None
(…) (…) (…) (…) (…) (…) (…) (…) (…) (…) (…) (…) (…) (…)

What are the data in each of these columns?

• unique_id: a unique integer to identify the row
• db_name: the name of the contigs database from which this data comes (only appears in output from multi-mode, in which multiple DBs are processed at once)
• genome_name/bin_name/contig_name: the identifier for the current sample, whether that is a genome, bin, or contig from a metagenome assembly
• kegg_module: the KEGG MODULE number for a metabolic pathway
• module_name/module_class/module_category/module_subcategory/module_definition: metabolic pathway information from the KEGG MODULE database
• module_completeness: a fraction between 0 and 1 indicating the proportion of steps in the metabolic pathway that have an associated KO annotation. To learn how this number is calculated, see the anvi-estimate-metabolism help page
• module_is_complete: a boolean value indicating whether the module_completeness score is above a certain threshold or not (the default threshold is 0.75)
• kofam_hits_in_module: a comma-separated list of the KO annotations that were found in the current sample and contribute to this metabolic pathway (these will be KOs from the metabolic pathway definition in the module_definition column)
• gene_caller_ids_in_module: a comma-separated list of the genes with KO annotations that contribute to this pathway, in the same order as the annotations in the kofam_hits_in_module column
• warnings: miscellaneous caveats to consider when interpreting the module_completeness score. For example, a warning like “No KOfam profile for K00172” would indicate that we cannot annotate K00172 because we have no HMM profile for that gene family, which means that any metabolic pathway containing this KO can never be fully complete (even if a gene from that family does exist in your sequences). Extra caution should be taken when considering the completeness of modules with warnings

Coverage and detection values in the output

If you use the flag --add-coverage and provide a profile database, additional columns containing coverage and detection data will be added for each sample in the profile database. Here is a mock example of the additional columns you will see (for a generic sample called ‘SAMPLE_1’):

SAMPLE_1_gene_coverages SAMPLE_1_avg_coverage SAMPLE_1_gene_detection SAMPLE_1_avg_detection
3.0,5.0,10.0,2.0 5.0 1.0,1.0,1.0,1.0 1.0

In this mock example, the module in this row has four gene calls in it. The SAMPLE_1_gene_coverages column lists the mean coverage of each of those genes in SAMPLE_1 (in the same order as the gene calls are listed in the gene_caller_ids_in_module column), and the SAMPLE_1_avg_coverage column holds the average of these values. As you probably expected, the detection columns are similarly defined, except that they contain detection values instead of coverage.

### ‘KOfam Hits in Modules’ Mode

The kofam_hits_in_modules output file will have the suffix kofam_hits_in_modules.txt. Each line in the file will represent information about one KOfam hit in a given genome, metagenome, or bin - but only KOs that are a part of at least one KEGG Module are included in this output. Hits are organized according to the KEGG module that they belong to, and more specifically the path through the KEGG module in which the KO appears.

What is a path through a KEGG module, you ask? Well. There is a lengthier explanation of this here, but we will go through it briefly below.

KEGG modules are metabolic pathways defined by a set of KOs. For example, here is the definition of module M00001, better known as “Glycolysis (Embden-Meyerhof pathway), glucose => pyruvate”:

(K00844,K12407,K00845,K00886,K08074,K00918) (K01810,K06859,K13810,K15916) (K00850,K16370,K21071,K00918) (K01623,K01624,K11645,K16305,K16306) K01803 ((K00134,K00150) K00927,K11389) (K01834,K15633,K15634,K15635) K01689 (K00873,K12406)

Spaces separate steps (reactions) in the metabolic pathway, and commas separate alternative KOs or alternative sub-pathways that can facilitate the same overall reaction. So a definition such as the one above can be “unrolled” into several different linear sequences of KOs, each of which we consider to be a possible “path” through the module. As an example, we can take the first option for every step in the Embden-Meyerhof pathway definition from above:

(K00844,K12407,K00845,K00886,K08074,K00918) (K01810,K06859,K13810,K15916) (K00850,K16370,K21071,K00918) (K01623,K01624,K11645,K16305,K16306) K01803 ((K00134,K00150) K00927,K11389) (K01834,K15633,K15634,K15635) K01689 (K00873,K12406)

to get the following path of KOs (which happens to be the path shown in the output example below):

K00844 K01810 K00850 K01623 K01803 K00134 K00927 K01834 K01689 K00873

For every KO in the path above that has a hit in the contigs-db, there will be a corresponding line in the ‘kofam_hits_in_modules’ output file. The same will occur for every possible path in every single KEGG module, resulting in a lot of lines and extremely repetitive but nicely parseable information.

Without further ado, here is an example of this output mode (also from the Infant Gut dataset):

unique_id db_name genome_name kofam_hit gene_caller_id contig path_id path path_completeness kegg_module module_completeness module_is_complete
0 E_faecalis_6240 Enterococcus_faecalis_6240 K01834 2342 Enterococcus_faecalis_6240_contig_00003_chromosome 0 K00844,K01810,K00850,K01623,K01803,K00134,K00927,K01834,K01689,K00873 0.8 Enterococcus_faecalis_6240 M00001 1.0 True
1 E_faecalis_6240 Enterococcus_faecalis_6240 K01834 2646 Enterococcus_faecalis_6240_contig_00003_chromosome 0 K00844,K01810,K00850,K01623,K01803,K00134,K00927,K01834,K01689,K00873 0.8 M00001 1.0 True
(…) (…) (…) (…) (…) (…) (…) (…) (…) (…) (…) (…)

Many of the columns in this data overlap with the modules_mode columns; you can find descriptions of those in the previous section. Below are the descriptions of new columns in this mode:

• kofam_hit: a KO annotation from the contigs database that contributes to a KEGG module
• gene_caller_id: the ID of the gene that is annotated with this KO
• contig: the contig on which this gene is present
• path_id: a unique identifier of the current path through the KEGG module
• path: the current path of KOs through the module (described above), which this KO annotation contributes to
• path_completeness: a fraction between 0 and 1 indicating the proportion of KOs in the current path that are annotated. To learn how this number is calculated, see the anvi-estimate-metabolism help page

Coverage and detection values in the output

The flag --add-coverage can also be used for this output mode, but in this case, the columns that are added (for each sample in the profile database) are slightly different. Here is a mock example:

SAMPLE_1_coverage SAMPLE_1_detection
3.0 1.0

Since each row is a single gene in this output mode, these columns will contain the coverage/detection values for that gene only.

### ‘KOfam Hits’ Mode

The kofam_hits output file will have the suffix kofam_hits.txt. Unlike the previous mode, this output will include ALL KO hits, regardless of whether the KO belongs to a KEGG Module or not. However, since only a subset of these KOs belong to modules, this output does not include module-related information like paths and module completeness.

Here is an example of this output mode (also from the Infant Gut dataset):

unique_id db_name genome_name ko gene_caller_id contig modules_with_ko definition
0 E_faecalis_6240 Enterococcus_faecalis_6240 K00845 1608 Enterococcus_faecalis_6240_contig_00003_chromosome M00001,M00549,M00892,M00909 glucokinase [EC:2.7.1.2]
1 E_faecalis_6240 Enterococcus_faecalis_6240 K01810 600 Enterococcus_faecalis_6240_contig_00003_chromosome M00001,M00004,M00114,M00892,M00909 glucose-6-phosphate isomerase [EC:5.3.1.9]
2 E_faecalis_6240 Enterococcus_faecalis_6240 K00850 225 Enterococcus_faecalis_6240_contig_00003_chromosome M00001,M00345 6-phosphofructokinase 1 [EC:2.7.1.11]
(…) (…) (…) (…) (…) (…) (…) (…)

Here are the descriptions of any new columns not yet discussed in the previous sections:

• ko: a KO that was annotated in the contigs database
• modules_with_ko: the KEGG modules (if any) that this KO belongs to
• definition: the function of this KO (typically the enzyme name and EC number)

Coverage and detection values in the output

If you use the flag --add-coverage and provide a profile database, you will get the same additional columns per row as described above in for kofam_hits_in_modules mode. That is, you will get one column per sample for coverage (containing the coverage value of the KO hit in the sample) and one column per sample for detection (containing the detection value of the KO hit in the sample).

### Custom Mode (for module data)

The modules_custom output mode will have user-defined content and the suffix modules_custom.txt (we currently only support output customization for modules data). See anvi-estimate-metabolism for an example command to work with this mode. The output file will look similar to the modules mode output, but with a different (sub)set of columns.

## Matrix format output

Matrix format is an output option when anvi-estimate-metabolism is working with multiple contigs databases at once. The purpose of this output type is to generate matrices of KEGG module statistics for easy visualization and clustering. Currently, the matrix-formatted output includes a module completeness matrix, a matrix of binary module presence/absence values, and a matrix of KO counts. In these matrices, each row is a KEGG module or KO, and each column is an input sample.

Here is an example of a module completeness matrix, for bins in a metagenome:

module bin_1 bin_2 bin_3 bin_4 bin_5 bin_6
M00001 1.00 0.00 1.00 1.00 1.00 0.00
M00002 1.00 0.00 1.00 1.00 1.00 1.00
M00003 0.88 0.00 1.00 0.75 1.00 0.88
M00004 0.88 0.00 0.88 0.88 0.88 0.00
M00005 1.00 0.00 1.00 1.00 1.00 1.00
(…) (…) (…) (…) (…) (…) (…)

Each cell of the matrix is the completeness score for the corresponding module in the corresponding sample (which is, in this case, a bin).

While the above is the default matrix format, some users may want to include more annotation information in the matrices so that it is easier to know what is going on when looking at the matrix data manually. You can add this metadata to the matrices by using the --include-metadata flag when running anvi-estimate-metabolism, and the output will look something like the following:

module module_name module_class module_category module_subcategory bin_1 bin_2 bin_3 bin_4 bin_5 bin_6
M00001 Glycolysis (Embden-Meyerhof pathway), glucose => pyruvate Pathway modules Carbohydrate metabolism Central carbohydrate metabolism 1.00 0.00 1.00 1.00 1.00 0.00
M00002 Glycolysis, core module involving three-carbon compounds Pathway modules Carbohydrate metabolism Central carbohydrate metabolism 1.00 0.00 1.00 1.00 1.00 1.00
M00003 Gluconeogenesis, oxaloacetate => fructose-6P Pathway modules Carbohydrate metabolism Central carbohydrate metabolism 0.88 0.00 1.00 0.75 1.00 0.88
(…) (…) (…) (…) (…) (…) (…)

The module completeness matrix files will have the suffix completeness-MATRIX.txt.

Module presence/absence matrix files will have the suffix presence-MATRIX.txt. In these files, each cell of the matrix will have either a 1.0 or a 0.0. A 1.0 indicates that the module has a completeness score above the module completeness threshold in that sample, while a 0.0 indicates that the module’s completeness score is not above the threshold.

Finally, KO hit matrix files will have the suffix ko_hits-MATRIX.txt. Each row of the matrix will be a KO, and each column will be an input sample. Cells in this matrix will contain an integer value, representing the number of times the KO was annotated in that sample.

Edit this file to update this information.