Generating Genomes
To generate a genome it is first necessary to simulate a Species Tree using simuLyon. And no, it is not possible to input an externally computed tree, but it will be in future versions
To simulate genomes, simuLyon starts with an ancestral genome at the root, with a given number of genes. For now in the current version, all genes families present in this genome have a single copy (so in this ancestral genome there are no duplicated genes).
A genome is an ordered collection of genes. So if we begin with a genome that has 5 genes, what we see is something like
| Position | Gene_family | Orientation | Id |
|---|---|---|---|
| 0 | 1 | + | 1 |
| 1 | 2 | - | 1 |
| 2 | 3 | + | 1 |
| 3 | 4 | + | 1 |
| 4 | 5 | - | 1 |
The meaning of this is:
- Position: The position in the genome. The genome is circular, so the position 4 is adjacent to the position 3 and 0
- Gene_family: The identifier of the gene family
- Orientation: The orientation of that gene in the genome
- Id: The identifier of the gene.
Genomes evolve undergoing a series of events:
- D: Duplications. A segment of the genome is duplicated. The new copy is inserted next to the old one
- L: Losses. A segment of the genome is lost
- T: Transfers. A segment of the genome is transferred to a contemporary species. The segment is inserted in a random position. Transfers can be replacement transfers
- C: Translocations. A segment of the genome changes its position within the genome
- I: Inversions. A segment of the genome inverts its position
- O: Originations. A new gene family appears and it is inserted in a random position
The rates in this case are genome-wise. For instance, a duplication rate of 3 means 3 duplication events per genome per unit of time.
There is also an additional rate for each event. This is called the extension_rate. This number (between 0 and 1) is the p parameters of a geometric distribution that controls the length of the affected segment.
For example, if DUPLICATION_EXTENSION == 1, the extension of the segment duplicated will be always 1 (meaning that only one gene is duplicated at a time)
By changing this parameter we can fine tune the extension associated to the different events. If inversions affect normally large chunks of the genome, it suffices to use a low p.
Origination of new gene families are always of size 1, meaning that it is not possible to have an origination of two gene families in the same step of time. Once that the full evolution of genomes has been simulated, simuLyon prints also the gene trees associated to the different gene families, all the events taking place in each gene family, the events taking place in each branch and the genomes of each node in the species tree.
There are two other events that do not depend intrinsically on genomes but in the species tree that is used to simulate genome evolution
- S: Speciation. When a genome arrives at a speciation node, the genome is divided and continues to evolve in both descendant branches
- E: Extinction. When a genome arrives at a extinction event, the genome stop its evolution
Some advances details regarding the genes identifiers: You might want to skip this part if you are reading this for the first time
Events that introduce nodes in the topology of the gene tree, change the identifier of the gene. For example, let us say that in the root we have a gene whose identifier is 1. If the genome where the gene undergoes a speciation, the two branches will inherit: one a gene whose identifier is 2 and the other one 3. A duplication will change also the identifiers of the duplicated genes. When a gene has been transferred, it changes the identifier of the gene remaining in the genome and in the recipient genome. This way is easy to track the events that have given rise to different tree topologies. Inversions and translocations do not introduce changes in the tree topology and for that reason they do not change the identifier of the affected genes.
In this model, genome event rates are specific for each branch of the species tree. Each time one new lineage emerges, a new value for its genome events rates its sampled from a user defined distribution (normal or lognormal), in which the mean corresponds to the value of the parent branch and the variance to the branch length
This model allows the user to fine control the genome rates
Genomes: A folder with one file per node of the species tree. Each file contains information about the genome composition.
Gene_families: A folder with one file per gene family. Each file contains information about the events taking place in that gene family. There are 3 fields.
1. Time: The time at which the event takes place 2. Event: The type of event that takes place in a given time (S, E, D, T, L, I, C, O and F. F stands for Final, meaning that the gene arrived alive till the end of the run) 3. Nodes: Some more information about the kind of event:
S, D and T: 6 fields separated by semicolons. This can be better understood looking at the picture:
- L, I, C, O and F: 2 fields separated by semicolons. First, the species tree branch where the event takes place and second, the identifier of the gene affected
GeneTrees: A folder containing the gene trees corresponding to the evolution of the different families and the gene trees pruned so that only surviving genes are represented. Please notice that gene trees with 2 or fewer surviving copies are simply not output. Mind that some families can appear at some point and then leave no surviving descendants!
There are two types of trees:
- _wholetree.nwk: A tree showing the whole evolution of that gene family
- _prunedtree.nwk: A tree in which the genes that have not survived till the present time have been removed. Normally you want to use this tree!
EventsPerBranch: (Not output by default) A folder with one file per branch of the species tree. Each file contains information about the events taking place in that branch. The codes are similar to the previously explained, but not the same. There are two main differences (for the sake of clarity). The first one is that transfers are divided into:
- LT: Leaving Transfers. Transfers that leave this branch
- AT: Arriving Transfers. Transfers that arrive at this branch.
The second difference is that the node of the nodes affected is given by:
GeneFamily_GeneIdentifier
So for example, if we go to the file n2_branchevents and we find the event L affecting at 4_3, means that the gene whose identifier is 3 belonging to the family 4 was lost in that branch in time given by the first column
Please also notice that in the case of events that affect to several genes, this will be reflected in the first column (several events taking place at the same unit of time)
Profiles: (Not output by default) Here there is a file called Profiles.tsv that contains the node of the species tree in the columns and the gene families in the rows. The entries give the number of copies that each gene family has for each node of the species tree.
DUPLICATION, TRANSFER, LOSS, INVERSION, TRANSLOCATION, ORIGINATION
The value for each type of event.
DUPLICATION_EXTENSION, TRANSFER_EXTENSION, LOSS_EXTENSION, INVERSION_EXTENSION, TRANSLOCATION_EXTENSION
The value of the p parameter of a geometric distribution that determines the extension of the genome (measured in number of genes) affected for event
REPLACEMENT_TRANSFER
A number between 0 and 1 controling the probability of replacement transfers (they only happen if there is a homologous position in the recipient genome)
STEM_FAMILIES
Number of gene families present in the ancestral genome at the root
MIN_GENOME_SIZE
The minimal size for a given genome. Smaller genomes will not be affected by losses events