# Splitting concatenated alignments by partition

To split an alignment, you need two input files: the concatenated alignment and the partition file. If you skip inputing partition file, the app will assume the partition is embedded in the alignment file.

The command is as below:

```Bash
segul split --input <path-to-alignment> --input-partition <path-to-partition-file>
```

For example:

```Bash
segul split --input concat-alignment.nexus --input-partition concat-alignment-partition.nex
```

## Supported partition format for splitting alignment

`segul` support two kind of RaXML format:

With datatype written in the partitions:

```Text
DNA, locus_1 = 1-5
DNA, locus_2 = 6-10
```

Or without the datatype:

```Text
locus_1 = 1-5
locus_2 = 6-10
```

For nexus partition:

```Text
#Nexus
begin set
  charset locus_1 = 1-5;
  charset locus_2 = 6-10;
end;
```

`segul` partition parser is robust when dealing with extra spaces, but forbid the use of spaces for the locus names.

This nexus format will work:

```Text
#Nexus
begin set
  charset locus_1 = 1 - 5 ;
  charset locus_2 = 6 - 10 ;
end;
```

But this will not work:

```Text
#Nexus
begin set
  charset 'locus 1' = 1-5;
  charset 'locus 2' = 6-10;
end;
```

Some other programs write locus names with file extension (contain dot). In that case, `segul` will replace it with underscore to avoid name conflict when changing file extension.

For example:

```Text
#Nexus
begin set
  charset locus_1.nex = 1 - 5 ;
  charset locus_2.nex = 6 - 10 ;
end;
```

In this case above, `segul` will write the output file as locus_1_nex.nex and locus_2_nex.nex (if save to nexus).