From e69efe348be65ae312b18d770f0662d09a28a667 Mon Sep 17 00:00:00 2001 From: alisandra Date: Mon, 13 Feb 2023 14:47:19 +0100 Subject: [PATCH] more docs and prep for v0.3.1 --- README.md | 15 ++++----------- setup.py | 2 +- 2 files changed, 5 insertions(+), 12 deletions(-) diff --git a/README.md b/README.md index f32621b..a4759b2 100644 --- a/README.md +++ b/README.md @@ -5,7 +5,6 @@ Gene calling with Deep Neural Networks. ## Disclaimer This software is undergoing active testing and development. -Build on it at your own risk. ## Goal Setup and train models for _de novo_ prediction of gene structure. @@ -46,16 +45,9 @@ This example focuses only on applying trained models for gene calling, only. Information on training and evaluating the models can be found in `docs`. ### Using trained models -> NOTE: the extensively evaluated models from the paper are available by -> running `git checkout v0.2.0` and following the instructions -> there in. But they were not yet _applicable_ for generating gff3 files. - -We are working towards training another round of models w/ the current -architecture. For now a preliminary land plant model is available and -will be used for the rest of the example. #### Acquire models -The best models for each or all lineages can automatically +The best models for each or all lineages can automatically be downloaded with the `fetch_helixer_models.py` script. The available lineages are `land_plant`, `vertebrate`, `invertebrate`, @@ -111,8 +103,9 @@ that generalize well to your target species. When in doubt selection via `--line this will use the best available model for that lineage. ##### `--subsequence-length` and overlapping parameters -> From v0.3.1 onwards these paramters are set to reasonable defaults when `--lineage` -> is used, but `--subsequence-length` will still need to be specified when using `--model-filepath`. +> From v0.3.1 onwards these parameters are set to reasonable defaults when `--lineage` +> is used, but `--subsequence-length` will still need to be specified when using `--model-filepath`, +> while the overlapping parameters can be derived automatically. Subsequence length controls how much of the genome the Neural Network can see at once, and should ideally be comfortably longer than the typical gene. diff --git a/setup.py b/setup.py index 036c249..3169a22 100644 --- a/setup.py +++ b/setup.py @@ -2,7 +2,7 @@ setup( name='helixer', - version='0.3.0', + version='0.3.1', description='Deep Learning fun on gene structure data', packages=['helixer', 'helixer.core', 'helixer.prediction', 'helixer.evaluation', 'helixer.tests', 'helixer.export'], package_data={'helixer': ['testdata/*.fa', 'testdata/*.gff']},