## GTDB taxonomy for mOTUs

We annotated all genomes used for mOTUs with GTDB-Tk. We then merge the annotation of the genomes at the mOTUs level, which represent clusters of genomes (see below for details).

Here is the taxonomy:

| mOTUs version          | File          | Annotation tool  |
| :--------------------: |:-------------:| :---------------:|
| mOTUs 3.0.0 - 3.0.3    | [mOTUs_3.0.0_GTDB_tax.tsv](https://sunagawalab.ethz.ch/share/MOTU_GTDB/mOTUs_3.0.0_GTDB_tax.tsv) | GTDB-Tk version 2.1 on database release [207](https://data.ace.uq.edu.au/public/gtdb/data/releases/release207/) (4.4 MB) |



## Annotation of mOTUs

Each mOTU cluster is composed of 1 or more genomes and for each genome we have a GTDB annotation that looks like:
```
GUT_GENOME002602	d__Bacteria;p__Bacteroidota;c__Bacteroidia;o__Bacteroidales;f__Bacteroidaceae;g__Bacteroides;s__Bacteroides fragilis
GUT_GENOME002402	d__Bacteria;p__Firmicutes;c__Bacilli;o__RF39;f__UBA660;g__;s__
```

Note that for each taxonomic level there is either an annotation (example: `s__Bacteroides fragilis`) or a missing annotation (example: `s__`). For the evaluation that we are doing here we consider `s__` (or `g__`, etc.) as `NA`.

Each taxonomic level in a mOTU can have three annotations:



### `Agreeing`

If at least 80% of the genomes agree to one annotation, then that annotation is selected. Note that we consider only annotations that are not `NA`. So for example, if we have 20 genomes in a mOTU, in all these cases there is an "agreeing" annotation:
```
Species:       s__Bacteroides fragilis
# of genomes:                       20
```
Annotated as `s__Bacteroides fragilis` as 100% of the genomes (20/20) agree at species level.


```
Species:       s__Bacteroides fragilis     NA
# of genomes:                       11      9
```
Annotated as `s__Bacteroides fragilis` as 100% of the genomes (11/11) agree at species level.

```
Species:       s__Bacteroides fragilis   s__Bacteroides vulgatus     NA
# of genomes:                       11                         1      8
```
Annotated as `s__Bacteroides fragilis` as 91.6% of the genomes (11/12) agree at species level.






### `Not annotated`

If all genomes at that taxonomic level do not have an annotation. Example:
```
Species:           NA
# of genomes:      20
```

Note that in the mOTUs taxonomy we report it as `Not_annotated [<last annotated level>]`. For example if a mOTUs is composed of 3 genomes with annotation:
```
GUT_GENOME002402	d__Bacteria;p__Firmicutes;c__Bacilli;o__RF39;         ;g__;s__
GUT_GENOME002403	d__Bacteria;p__Firmicutes;c__Bacilli;o__RF39;f__UBA660;g__;s__
GUT_GENOME002404	d__Bacteria;p__Firmicutes;c__Bacilli;o__RF39;f__UBA660;g__;s__
```

The mOTU annotation will be:
```
ref_mOTU_v3_00002	d__Bacteria;p__Firmicutes;c__Bacilli;o__RF39;f__UBA660;Not_annotated [f__UBA660];Not_annotated [f__UBA660]
```







### `Incongruent`
If the genomes do not agree (<80% agreement) at one specific taxonomic level. Example:

```
Species:       s__Bacteroides fragilis   s__Bacteroides vulgatus     NA
# of genomes:                       11                         7      2
```

Here the one with the highest agreement is `s__Bacteroides fragilis`, but only 11 out of 18 (11+7, note that we do not count the `NA`), which is 61% (below 80%), agree. Hence this level will be annotated as `Incongruent [<last annotated level>]`.

Here is an example with a mOTUs with 5 genomes:
```
GUT_GENOME002602  d__Bacteria;p__Bacteroidota;c__Bacteroidia;o__Bacteroidales;f__Bacteroidaceae;g__Bacteroides;s__Bacteroides fragilis
GUT_GENOME002603  d__Bacteria;p__Bacteroidota;c__Bacteroidia;o__Bacteroidales;f__Bacteroidaceae;g__Bacteroides;s__Bacteroides fragilis
GUT_GENOME002604  d__Bacteria;p__Bacteroidota;c__Bacteroidia;o__Bacteroidales;f__Bacteroidaceae;g__Bacteroides;s__Bacteroides vulgatus
GUT_GENOME002605  d__Bacteria;p__Bacteroidota;c__Bacteroidia;o__Bacteroidales;f__Bacteroidaceae;g__Bacteroides;s__Bacteroides vulgatus
GUT_GENOME002606  d__Bacteria;p__Bacteroidota;c__Bacteroidia;o__Bacteroidales;f__Bacteroidaceae;g__Bacteroides;s__Bacteroides vulgatus
```

Where at species level we have 3 `s__Bacteroides vulgatus` and 2 `s__Bacteroides fragilis`. The mOTUs annotation is:
```
d__Bacteria;p__Bacteroidota;c__Bacteroidia;o__Bacteroidales;f__Bacteroidaceae;g__Bacteroides;Incongruent [g__Bacteroides]
```

Note: when a level is incongruent, all levels underneath are set to incongruent. If we don't do this we can have a situation where from phylum to level it is incongruent and at species level it is incongruent (if some annotations are NA at species level). For example say we have a mOTUs with two genomes:
```
{'d__Bacteria': 2}
{'p__Bacteroidota': 1, 'p__Riflebacteria': 1}
{'c__Bacteroidia': 1, 'c__Ozemobacteria': 1}
{'o__Bacteroidales': 1, 'o__Ozemobacterales': 1}
{'f__Bacteroidaceae': 1, 'f__Ozemobacteraceae': 1}
{'g__Prevotella': 1, 'g__RUG334': 1}
{'NA': 1, 's__RUG334': 1}
```

One of the two genomes is not annotated at species level, hence `s__RUG334` would have a 100% agreement and it would not be "Incongruent" like the genus level. But we prevent this, hence the mOTUs annotation is:
```
ext_mOTU_v3_22969	d__Bacteria	Incongruent [d__Bacteria]	Incongruent [d__Bacteria]	Incongruent [d__Bacteria]	Incongruent [d__Bacteria]	Incongruent [d__Bacteria]	Incongruent [d__Bacteria]
```