Running ORFold:#

Advanced run:#

Mapping of the fold potential and the disorder and aggregation propensities along the genome of an organism#

In the previous section we presented how to launch ORFold on a set of amino acid sequences stored in a fasta file. However, the originality of ORFold relies on the fact that the user can manually inspect the distribution of the properties estimated with ORFold (fold potential, and disorder and aggregation propensities) along a genome of interest. In this case, the user must provide the genome annotation file (gff) along with the input fasta file. ORFold will return new gff files (one per studied property) that contain for the ORFs provided in the input fasta file, their corresponding property scores (fold potential, disorder or aggregation propensities). The values are stored in the column #9 of the output gff files. The gff files can be subsequently uploaded on a genome viewer such as IGV [1].

The input gff file must be given with the -gff option as follows:

orfold -faa /database/sequences.fasta -options HIT -gff /database/sequences.gff

ORFold generates a sequences.tab file containing the fold potential, and the disorder and aggregation propensities of each sequence present in the input fasta file. Additionally, ORFold produces three new gff files:

sequences_HCA.gff
sequences_IUPRED.gff
sequences_TANGO.gff

The output gff files are identical to the one provided by the user except that for the sequences present at the same time in the fasta and the gff files that were given as inputs, the column #9 is now replaced by the fold potential, or the disorder or aggregation propensities calculated by ORFold. That way, the corresponding sequences can be colored according to these values on a genome viewer, thereby enabling the visual inspection of these properties along the input genome. Notice that on IGV, blue indicates low values (for all mapped properties) while red indicates high values.

HCA Scale
Figure 1: Color scale for the HCA score values

Note

Notice that the ID of the sequences given in the fasta file (i.e. annotation after the ">" in the fasta file) must be strictly identical to those of the corresponding sequences in the gff file (i.e. ID indicated in the column #3 of the gff file).

Dealing with multiple files at the same time:#

ORFold can handle multiple input files (fasta and gff) and will associate the fasta and gff files according to the following rules:

If the user provides the same number of fasta and gff files, ORFold associates them based on their root name, no matter the order of the files.
```
orfold -faa /database/sequences_Y.fasta /database/sequences_X.fasta -options H -gff /database/sequences_X.gff /database/sequences_Y.gff
```
In this case, ORFold associates:
- sequence_Y.fasta with sequence_Y.gff
- sequence_X.fasta with sequence_X.gff
I If the user provides the same number of fasta and gff files, but their root names are not identical, ORFold associates them according to the order of the files in the command line.
```
orfold -faa /database/sequences_Y.fasta /database/sequences_X.fasta -options H -gff /database/sequences_A.gff /database/sequences_B.gff
```
In this case, ORFold associates:
- sequence_Y.fasta with sequence_A.gff
- sequence_X.fasta with sequence_B.gff
If the user provides multiple fasta files and only one gff file, then:
- If the name of the gff file matches with the name of one of the fasta files, the two files are associated, while the other fasta files are not associated to the input gff file.
```
orfold -faa /database/sequences_Y.fasta /database/sequences_X.fasta -options H -gff /database/sequences_X.gff
```
  ORFold associates:
  - sequences_X.fasta with sequences_X.gff
  - sequences_Y.fasta with nothing
- If the name of the gff file does not match with any of the names of the fasta files, then the gff file will be associated with all the fasta files, considering that the fasta files correspond to different subgroups of the same dataset.
```
orfold -faa /database/sequences_Y.fasta /database/sequences_X.fasta -options H -gff /database/sequences_B.gff
```
  ORFold associates:
  - sequences_Y.fasta with sequences_B.gff
  - sequences_X.fasta with sequences_B.gff

If the user provides multiple fasta and gff files (but not the same number), all gff files must have a corresponding fasta file with the same root name. Otherwise, ORFold will give an ERROR message.

orfold -faa /database/sequences_Y.fasta /database/sequences_X.fasta /database/sequences_Z.fasta -options H -gff /database/sequences_Z.fasta /database/sequences_Y.gff

ORFold associates:

sequences_Y.fasta with sequences_Y.gff
sequences_Z.fasta with sequences_Z.gff
sequences_X.fasta with nothing

orfold -faa /database/sequences_Y.fasta /database/sequences_X.fasta /database/sequences_Z.fasta -options H -gff /database/sequences_B.fasta /database/sequences_A.gff

ORFold will give the following error message:

Oups! You provided GFF file(s) which has/have no correspondance to the input FASTA files

All these examples assume that the input files are stored in the /database/ directory of the container. All the outputs are written in the /workdir/orfold/ directory of the container.

Running ORFold on subsets of randomly selected sequences#

Working with complete genomes could generate big amounts of sequences which can dramatically increase the computational time of ORFold when dealing with large genomes (especially if the estimation of the disorder and aggregation propensities are activated). If the user does not want to treat all the sequences, he/she can create a random sample of the input sequences, large enough to have an estimation of the distribution of the studied properties of its dataset from a representative sample of the input sequences. To do so, the user must indicate the number of sequences that are to be randomly selected with the -N option. For a representative dataset, we recommend selecting at least 10000 sequences.

orfold -faa /database/sequences.fasta -options HIT -gff /database/sequences.gff -N 10000

In this example, ORFold will estimate the fold potential, and the disorder and aggregation propensities on a sample of 10000 sequences extracted randomly from the initial sequences.fasta file.

Note

If the user works with more than one fasta file and wishes to create random samples for all the input sequence files, he/she has to indicate in the -N option the size for each input file explicitly (in the same order as the inputs passed in the -fna option).

orfold -faa /database/sequences_X.pfasta /database/sequences_Y.pfasta -options H -N 1500 3000

Also, if the user wants to sample two subsets of same sizes, he/she has to indicate the subset sizes explicitly for each input

orfold -faa /database/sequences_X.pfasta /database/sequences_Y.pfasta -options H -N 1500 1500

If the user whishes to calculate the fold potential of all the sequences of one of the given inputs, he has to indicate it with the "all" flag (again with respect to the order of input files)

orfold -faa /database/sequences_X.pfasta /database/sequences_Y.pfasta -options H -N all 3000

In this case, ORFold will calculate the fold potential for all the sequences in the sequences_X file while will generate a random sample of 3000 sequences for the sequences_Y file.

References

Robinson JT, Thorvaldsdóttir H, Winckler W, et al (2011) Integrative genomics viewer. Nature biotechnology 29:24–26