endrebak

import copy

tmp_path = '/nethome/jkpickrell/1kG_data/covariance_matrix/'
# tmp_path_EUR = tmp_path
# tmp_path_ASN = '/nethome/jkpickrell/1kG_data/covariance_matrix/ASN/'
# tmp_path_AFR = '/nethome/jkpickrell/1kG_data/covariance_matrix/AFR/'

def return_conf(path):
	return {
		'partition_root' : path,

endrebak

39967768 39967768 0.249157622127387
39967768 39967824 0.022333456325143934
39967768 39967950 -0.00434938492491698
39967768 39968210 -0.01366949547831051
39967768 39968231 -0.019882902513906196
39967768 39968596 -0.007456088442714824
39967768 39968633 0.00348960255080374
39967768 39968753 0.0006979205101607479
39967768 39969173 0.0006979205101607479
39967768 39969251 0.18011222422567374

endrebak

rule compute_kde:
    input:
        "{prefix}/data/{genome}/HMM_states/{statistic}/correlations_all.gz"
    output:
        "{prefix}/data/{genome}/HMM_states/{statistic}/cutoff.txt"
    run:
        f = input[0]
        o = output[0]

        from scipy.stats import gaussian_kde

endrebak

@cython.boundscheck(False)
@cython.wraparound(False)
@cython.initializedcheck(False)
cpdef _getitem(const long [::1] runs, const double [::1] values, const long [::1] run_cumsum, start, end):

    cdef:
        int i = 0
        int arr_length = 100
        int nfound = 0
        # int foundsum = 0

endrebak

rule gene_biotypes:
    input:
        regions = gene_biotype_infiles,
        annotation = "{prefix}/data/{genome}/annotation.tsv"
    output:
        "{prefix}/data/{genome}/{hmm_or_anatomy}_regions/{cutoff}/gene_biotype_counts.tsv"
    run:
        df = pd.read_table(input.regions, header=0)
        df2 = pd.read_csv(input.annotation, header=0, sep=",")

endrebak

import pyranges_db as db
import pyranges as pr

gr = db.gencode.genes("human") # takes a while to download from ftp
# Wall time: 2min 1s

gr.to_gtf("gencode_human.gtf.gz") # takes a while to gzip and write to disk
# Wall time: 4min 3s

# subset for faster operations

endrebak

#!/usr/bin/env python3
#cython: language_level=3, boundscheck=False, wraparound=False, initializedcheck=False, cdivision=True

import sys, math, gzip
import numpy as np
# cimport numpy as cnp
import pandas as pd

from time import time

endrebak

#!/usr/bin/env python3
#cython: language_level=3, boundscheck=False, wraparound=False, initializedcheck=False, cdivision=True

import sys, math, gzip
import numpy as np
import pandas as pd

from time import time

from libc.math cimport exp, fabs

endrebak

#!/usr/bin/env python3
#cython: language_level=3, boundscheck=False, wraparound=False, initializedcheck=False, cdivision=True

import sys, math, gzip
import numpy as np
import pandas as pd

from time import time

from libc.math cimport exp, fabs

endrebak

half_window_size = 3
len_haps = 10

# half_window_size += 1

for i in range(0, half_window_size):

        for j in range(i):

            j1 = i - j