calbors commited on
Commit
e7d85fe
·
verified ·
1 Parent(s): 2aa27b1

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +2 -0
README.md CHANGED
@@ -3,6 +3,8 @@ license: cc
3
  ---
4
  PhyloGPN is a convolutional neural network that takes encoded DNA sequences as input and outputs rate matrix parameters for [Felsenstein's 1981 model](https://en.wikipedia.org/wiki/Models_of_DNA_evolution#F81_model_(Felsenstein_1981)) (the F81 model, for short). It was trained to maximize the likelihood of columns in the [Zoonomia alignment](https://cglgenomics.ucsc.edu/november-2023-nature-zoonomia-with-expanded-primates-alignment/) given a phylogenetic tree. The stationary distribution of the substitution process described by the F81 model indicates the relative viability of each allele at any given locus. As a result, PhyloGPN is formally a genomic language model. It can be used for transfer learning.
5
 
 
 
6
  The following Python snippet shows how to obtain embeddings and log rate parameters from PhyloGPN for each site in a batch of sequences. Note that PhyloGPN is designed as a sliding window function: it takes a batch of \\(b\\) encoded sequences of any given length \\(\ell >= 481\\) as input and yields outputs for the \\(b \times (\ell - 480)\\) central positions.
7
 
8
  ```python
 
3
  ---
4
  PhyloGPN is a convolutional neural network that takes encoded DNA sequences as input and outputs rate matrix parameters for [Felsenstein's 1981 model](https://en.wikipedia.org/wiki/Models_of_DNA_evolution#F81_model_(Felsenstein_1981)) (the F81 model, for short). It was trained to maximize the likelihood of columns in the [Zoonomia alignment](https://cglgenomics.ucsc.edu/november-2023-nature-zoonomia-with-expanded-primates-alignment/) given a phylogenetic tree. The stationary distribution of the substitution process described by the F81 model indicates the relative viability of each allele at any given locus. As a result, PhyloGPN is formally a genomic language model. It can be used for transfer learning.
5
 
6
+ A more detailed description of the model is available [here, in the preprint](https://arxiv.org/abs/2503.03773).
7
+
8
  The following Python snippet shows how to obtain embeddings and log rate parameters from PhyloGPN for each site in a batch of sequences. Note that PhyloGPN is designed as a sliding window function: it takes a batch of \\(b\\) encoded sequences of any given length \\(\ell >= 481\\) as input and yields outputs for the \\(b \times (\ell - 480)\\) central positions.
9
 
10
  ```python