itrails 0.1.0

Tree reconstruction of ancestry using incomplete lineage sorting

iTRAILS

Software

iTRAILS is a framework to jointly model incomplete lineage sorting (ILS) and gene flow (GF). iTRAILS uses the multispecies coalescent to approximate the ancestral recombination history of three populations and an outgroup. In brief, iTRAILS infers the ancestral tree from a multiple genome alignment, unbiasedly estimating ancestral effective population sizes, speciation times, and introgression time and proportions. iTRAILS also performs posterior decoding.

iTRAILS can be installed as both a PyPi and a conda package, and it's codebase is available at:

https://github.com/trails-phylogeny/itrails

Once the package is installed, iTRAILS is initiaited and controlled through 2 main console commands.

Optimization of Parameters

iTRAILS estimates key demographic parameters—such as speciation times, ancestral effective population sizes (Ne), and recombination rate (ρ)—by maximizing the log-likelihood computed from the coalescent HMM. iTRAILS takes into account the following parameters:

• Population size parameters:
  • N_AB: Effective size of the ancestral population between the two most closely related species.
  • N_ABC: Effective size of the ancestral population between the three species.
• Time parameters (generations):
  • T_1: Assuming ultrametric model, time from sample to coalescence of species A and B.
  • T_2: Time between coalescence of species A and B and coalescence of AB ancestral with species C.
  • T_3: Time between coalescence of species ABC and coalescence of ABC ancestral with outgroup species.
  • T_A: Time from sample of species A until coalescence with species B.
  • T_B: Time from sample of species B until coalescence with species A.
  • T_C: Time from sample of species C until coalescence with AB ancestral.
  • T_Upper: Time from start of last discretized time interval in ABC ancestral until coalescence with outgroup.
  • T_Out: Time from sample of outgorup until coalescence with ABC ancestral.
• Other parameters:
  • Recombination rate (ρ): NUmber of recombinations per site per generation.
  • Mutation rate (μ): Number of mutations per site per generation.

Not all of the previously described parameters are independent from each other. Therefore, definition of parameters should comply with the following restrictions:

• T_2, N_AB, N_ABC, ρ must always be defined as a fixed parameter os as a parameter to optimize.
• μ must always be defined as a fixed parameter.
• T_Upper and T_3: At least one of them should be defined (fixed or optimized). If only one of them is defined, the other one will be automatically calculated.
• T_out: Can be defined or omitted, if defined it must be fixed, if omitted it will be automatically calculated.
• T_A, T_B, T_C and T_1 should be defined (fixed or optimized) in the following combinations, the non defined parameters will take values as defined in the following table:

Specified parameters	Used T_A	Used T_B	Used T_C
T_A / T_B / T_C	T_A	T_B	T_C
T_1 / T_A	T_A	T_1	T_1 + T_2
T_1 / T_B	T_1	T_B	T_1 + T_2
T_1 / T_C	T_1	T_1	T_C
T_A / T_B	T_A	T_B	$\frac{T_A + T_B}{2} + T_2$
T_A / T_C	T_A	$\frac{T_A + T_C - T_2}{2}$	T_C
T_B / T_C	$\frac{T_B + T_C - T_2}{2}$	T_B	T_C
T_1	T_1	T_1	T_1 + T_2

In order to run an optimization, iTRAILS needs two input files.

• A four-way genome alignment (MAF file) between three species and an outgroup.

$ cat alignment.maf

##maf version=1 scoring=none
# generated by Biopython

a score=3.25073e+06
s gorGor5.CYUI01015145v1                           8162515  3382 +         8455952 tcaaaaaactatttcttgagcattcattaagtgcaaa
s hg38.chr1                                       24998515  3414 +       248956422 TCAAAAAACTATTTCTTGAGCATTCATTAAGTGCAAA
s panTro5.chr1                                    24190887  3415 +       228573443 tcaaaaaactatctcttgagcattcattaagtgcaaa
s ponAbe2.chr1                                    24531025  3415 -       229942017 tcaaaaaactctttcttgagcattcattaagtgcaaa

a score=4.56614e+06

• A configuration file (.yaml) that specifies fixed parameters, parameters to optimize, as well as the alignment file path (optional) and output path (optional). An example configuration file is found in the package and can be seen in console with the following function: