Skip to content

Latest commit

 

History

History
45 lines (31 loc) · 1.31 KB

README.md

File metadata and controls

45 lines (31 loc) · 1.31 KB
Raw

Deep Variant

Introduction

Implementation of Finding candidate variants and Creating images around candidate variants sections (SNP only so far) in Google Variant Caller Paper.

Run run.sh to find candidate variants, then generate 3 types of images which are ref, het and hom-alt for further CNN training network (using TensorFlow) by given a specific range.

Structure

  • find_candidate
    • find.py
    • find_candidate.sh
  • image_generation
    • draw.py
    • gen_image.py
  • label_classification
    • label_classification.py
  • tools
    • samtools-1.5/
    • image_count.sh
    • run_sample.sh
  • run.sh

find_candidate.sh - select the following 1000000 position to find candidate.

find.py - find candidate variants.

draw.py - draw images with feature.

gen_image.py - preprocessing of image drawing.

label_classification.py classify those images into ref, het and hom-alt then call image generation script.

samtools1.5/ - samtools.

image_count.sh - count image in each class.

run_sample.sh - sample script for understanding how to run.

Usage

Usage: run.sh <chr> <start pos> <end pos>

sample run script can be found in tools/run_sample.sh.