Phlya

# Fastqs can be provided as:
# -- a pairs of relative/absolute paths
# -- sra:<SRA_NUMBER>, optionally followed by the indices of the first and
# the last entry in the SRA in the form of "?start=<first>&end=<last>
# [to implement] -- as a path to a folder with fastqs '<base_folder>', with the structure
# <base_folder>/<library_name>/<run_name>/, with each folder containing only
# two fastq.gz files
input:
    raw_reads_paths:
# Hi-C

Phlya

# Fastqs can be provided as:
# -- a pairs of relative/absolute paths
# -- sra:<SRA_NUMBER>, optionally followed by the indices of the first and
# the last entry in the SRA in the form of "?start=<first>&end=<last>
# [to implement] -- as a path to a folder with fastqs '<base_folder>', with the structure 
# <base_folder>/<library_name>/<run_name>/, with each folder containing only
# two fastq.gz files
input:
    raw_reads_paths:
        Schmitt2016_STL002_Adrenal-1:

Phlya

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Thu Jul 18 10:07:49 2019

@author: s1529682
"""

from cooltools.lib import numutils
import numpy as np

Phlya

Phlya

def fetch_random_means(clr1, exp1, clr2, exp2, region_left, region_right=None, n=1000): 
    assert clr1.binsize==clr2.binsize
    if region_right is None:
        region_right = region_left
    
    chrom_left, start_left, end_left = cooler.util.parse_region_string(region_left)
    chrom_right, start_right, end_right = cooler.util.parse_region_string(region_right)
    assert chrom_left == chrom_right

    bin1_left = start_left//clr1.binsize

Phlya

def pcolormesh_45deg(matrix_c, ax, start=0, resolution=1, *args, **kwargs):
    start_pos_vector = [start+resolution*i for i in range(len(matrix_c)+1)]
    import itertools
    n = matrix_c.shape[0]
    t = np.array([[1, 0.5], [-1, 0.5]])
    matrix_a = np.dot(np.array([(i[1], i[0])
                                for i in itertools.product(start_pos_vector[::-1],
                                                           start_pos_vector)]), t)
    x = matrix_a[:, 1].reshape(n + 1, n + 1)
    y = matrix_a[:, 0].reshape(n + 1, n + 1)

Phlya

### do required imports
### make a dict with coolers, like this {'name':cooler.Cooler('path'), ...}
### make a dict with coordinates, smth like this: {'300kb_1': ('chr6', 80200000, 84200000), ...}


# Here is the key function to plot heatmaps like triangles
def pcolormesh_45deg(matrix_c, ax, start=0, resolution=1, *args, **kwargs):
    start_pos_vector = [start+resolution*i for i in range(len(matrix_c)+1)]
    import itertools
    n = matrix_c.shape[0]

Phlya

def dedup(dots, hiccups_filter=True):
    newdots = []
    ress = list(sorted(set(dots['res'])))
    for chrom in sorted(set(dots['chrom1'])):
        chromdots = dots[dots['chrom1']==chrom].sort_values(['start1', 'start2']).reset_index(drop=True)
        for res in ress:
            chromdots['Supported_%s' % res] = (chromdots['res']==res)
        tree = spatial.cKDTree(chromdots[['start1', 'start2']]) #Not sure what's the best coordinate to use, of the strongest pixel, centroid, or the center of the dot?
        drop = []
        for i, j in tree.query_pairs(r=50000):

Phlya

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Tue Sep  4 11:54:20 2018

@author: Ilya Flyamer
"""

import pandas as pd
import numpy as np

Phlya

process{

    // SGE node config
    executor = 'sge'
    penv = 'sharedmem'
    clusterOptions = '-l h_vmem=4G -V'
    time='12h'
    cpus = 1
    maxRetries = 2
    distillerTmpDir='/exports/eddie/scratch/ifliamer'