生信4--用python将blasttab+格式文件写成class对象并给予过滤方法

blasttab+结果写成class

先每一行解析一下

class LastHit(object):
    def __init__(self,line):
        fields = line.rstrip().split("\t")
        self.query_sequence_id = fields[0]
        self.subject_sequence_id = fields[1]
        self.identity = float(fields[2])
        self.alignment_length = int(fields[3])
        self.mismatchs = int(fields[4])
        self.gaps = int(fields[5])
        self.qstart = int(fields[6])            # qstart: start of alignment in query
        self.qend = int(fields[7])              # qend: end of alignment in query
        self.sstart = int(fields[8])            # sstart: start of alignment in subject
        self.send = int(fields[9])              # send: end of alignment in subject
        self.evalue = float(fields[10])         # evalue: expect value
        self.score = float(fields[11])
        self.query_length = int(fields[12])
        self.subject_length = int(fields[13])
        self.undientify = int(fields[14])
        
        if self.qstart >= self.qend:
            tmp_position = self.qstart
            self.qstart = self.qend
            self.qend = tmp_position

发现不是很方便，那就把每一行形成一个列表写成类对象，用了pandas模块方便一点

class LastHits(object):
   def __init__(self,last_result_list):
       self.last_hits = []
       for i in last_result_list:
           self.last_hits.append(LastHit(i))
       
       self.last_result_pd_list = []
       for i in last_result_list:
           self.last_result_pd_list.append(i.split("\t"))
       
       columns_names = ['query_sequence_id','subject_sequence_id','identity','alignment_length','mismatchs','gaps','qstart','qend','sstart','send','evalue','score','query_length','subject_length','unidentify']
       self.last_result_pd = pd.DataFrame(self.last_result_pd_list, columns=columns_names, dtype=float)
       
   def filter_result_pd(min_identity = 90,min_alignment_length = 400):
       if min_identity is not None:
           self.last_result_pd[self.last_result_pd['identity'] > 90]
       if min_alignment_length is not None:
           self.last_result_pd[self.last_result_pd['alignment_length'] > 400]

这样，运行下面代码

cmd = 'lastal -Q1 -P 4 -q 1 -b 1 -Q 0 -a 1 -e 45 -f BlastTab+ AMR.db {}'.format(input_fastq_file) #可以改变Last得分矩阵的参数
cmd += " | grep -v ^# -"
    
last_hits = []
f = os.popen(cmd)
last_result_list = f.read().rstrip().split("\n") #列表，一行为一元素

然后把结果解析成LastHits对象

last_hits = LastHits(last_result_list)

LastHits对象就拥有了last_hits、last_result_pd_list、last_result_pd属性，并且拥有了filter_result_pd（）过滤方法，后面就是不断添加方法

这种思路就是python面对对象的编程思路，我就做个简单演示