spark kmer计算

• 输入文件：fa格式的文件
• 输出结果：kmer的频数和对应的kmer类型和计数
  1.将fq.gz的文件转换成fa文件：

#!/usr/bin/python env
# -*- coding:utf-8 -*-
import os
import re
import os.path
import gzip
import sys

#在这里可以写一个函数用来将文件转换成id和序列对应的字典
#需要用到哪个转化操作呢？考虑先尝试使用filter或者map

'''
@r261DFRAAXX100204:1:100:10494:3070/1
ACTGCATCCTGGAAAGAATCAATGGTGGCCGGAAAGTGTTTTTCAAATACAAGAGTGACAATGTGCCCTGTTGTTT
+
ACCCCCCCCCCCCCCCCCCCCCCCCCCCCCBC?CCCCCCCCC@@CACCCCCACCCCCCCCCCCCCCCCCCCCCCCC
'''

#这里是利用python直接读取压缩的fastq文件
def read_gz_file(path):
    if os.path.exists(path):
        with gzip.open(path,'rt') as pf:
            for line in pf:
                yield line
    else:
        print 'the path [{}] is not exist!'.format(path)

def ReadFastq(fastq):
    flag = 1
    dict_fq={}
    if fastq.endswith('gz'):
        con = read_gz_file(fastq)
        if getattr(con,'__iter__','None'):
            for line in con:
                line=line.strip()
                flag_index = flag%4
                if flag_index == 1:
                    id = line
                if flag%4 == 2:
                    seq = line
                else:
                    flag +=1
                    continue
                dict_fq[id] = seq
                flag+=1
            return dict_fq
    else:
        with open (fastq,'r') as fqr:
            for line in fqr.readlines():
                line = line.strip()
                flag_index = flag%4
                if flag_index == 1:
                    id = line
                if flag%4 == 2:
                    seq = line
                else:
                    flag +=1
                    continue
                dict_fq[id] = seq
                flag+=1
            return dict_fq

def convert_to_fa(dict_hash,output):
    with open (output,'w') as fr:
        for i in dict_hash.keys():
            fr.write(i+'\n')
            fr.write(dict_hash[i]+'\n')

if __name__ == '__main__':
    input = sys.argv[1]
    output = sys.argv[2]

    dic_fa = ReadFastq(input)
    convert_to_fa(dic_fa,output)

2.将reads打断成kmer并统计kmer的频数

#!/usr/bin/env python
# coding=utf-8
import os
import sys
import re
from pyspark import SparkConf, SparkContext

input_fasta_file ='/home/yueyao/Spark/00.data/both.fa'

conf = SparkConf().setMaster("local").setAppName("Yue Yao app")
sc = SparkContext(conf = conf)
fasta_file = sc.textFile(input_fasta_file)

#这里是对fasta文件进行转化操作，过滤掉reads的名称
reads_fa = fasta_file.filter(lambda line :">" not in line)

#这个函数用来将reads打断成kmer，这里的kmer是25,返回一个列表
def map_file(line):
    seq_lis=[]
    for i in range(len(line)-25+1):
        sub_seq = line[i:i+25]
        seq_lis.append(sub_seq)
    return seq_lis

kmer_list = reads_fa.flatMap(map_file)
#对打断的kmer进行计数
kmer_count = kmer_list.map(lambda id:(id,1))
kmer_total_count = kmer_count.reduceByKey(lambda a,b:(a+b))
#这里过滤掉了含有N的kmer
kmer_not_contain_N = kmer_total_count.filter(lambda line :"N" not in line[0])
kmer_key=kmer_not_contain_N.keys()
#统计kmer的种类，并计数
kmer_vari_count = kmer_not_contain_N.map(lambda kmer_vari:(kmer_vari[1],1))
kmer_histo = kmer_vari_count.reduceByKey(lambda a,b:(a+b))
#输出kmer频数的结果
kmer_histo.saveAsTextFile('Kmer25.histo')
kmer_not_contain_N.saveAsTextFile('kmer25')
kmer_key.saveAsTextFile('kmer25_key')