按群集拆分bam，然后使用checkpoin按群集合并bam

Question

我有三个来自3个不同样本的单单元bam文件，需要通过集群将它们拆分为更小的bam。然后，我需要为相同的集群合并来自不同示例的bam文件。我试过使用检查站，但有点迷路了。https://snakemake.readthedocs.io/en/stable/snakefiles/rules.html

这是我发布的这个问题的延续split bam files to (variable) pre-defined number of small bam files depending on the sample

SAMPLE_cluster = { "SampleA" : [ "1", "2", "3" ], "SampleB" :  [ "1" ], "SampleC" : [ "1", "2" ] }

CLUSTERS = []
for sample in SAMPLE_cluster:
    CLUSTERS.extend(SAMPLE_cluster[sample])
CLUSTERS = sorted(set(CLUSTERS)

rule all:
    input: expand("01merged_bam/{cluster_id}.bam, cluster_id = CLUSTERS)

checkpoint split_bam:
    input: "{sample}.bam"
    output: directory("01split_bam/{sample}/")
    shell:
       """
       split_bam.sh {input} 
       """
## the split_bam.sh will split the bam file to "01split_bam/{sample}/{sample}_{cluster_id}.bam" 

def merge_bam_input(wildcards):
    checkpoint_output = checkpoints.split_bam.get(**wildcards).output[0]
    return expand("01split_bam/{sample}/{sample}_{{cluster_id}}.bam", \
                sample = glob_wildcards(os.path.join(checkpoint_output, "{sample}_{cluster_id}.bam")).sample)


rule merge_bam_per_cluster:
    input: merge_bam_input
    output: "01merged_bam/{cluster_id}.bam"
    log: "00log/{cluster_id}.merge_bam.log"
    threads: 2
    shell:
        """
        samtools merge -@ 2 -r {output} {input}
        """

根据群集编号，每个群集的规则合并\u bam\u的输入将更改：

例如，对于集群1：“01split\u bam/SampleC/SampleC\u 1.bam”、“01split\u bam/SampleB/SampleB\u 1.bam”、“01split\u bam/SampleC/SampleC\u 1.bam”。你知道吗

对于集群2：“01split\u bam/samdeace/samdeace\u 2.bam”、“01split\u bam/SampleC/SampleC\u 2.bam”。你知道吗

对于集群3：“01split\u bam/samdeace/samdeace\u 3.bam”。你知道吗

Answer 1

我决定不使用checkpoint，而是使用input函数来获取

SAMPLE_cluster = { "SampleA" : [ "1", "2", "3" ], "SampleB" :  [ "1" ], "SampleC" : [ "1", "2" ] }

# reverse the mapping
cluster_sample = {'1':['sampleA','sample'B','sampleC'], '2':['sampleA', 'sampleC'], '3':['sampleA']}

rule split_bam:
    input: "{sample}.bam"
    output: "split.touch"
    shell:
       """
       split_bam {input} 
       touch split.touch
       """
rule index_split_bam:
    input: "split.touch"
    output: "split_bam/{sample}_{cluster_id}.bam.bai"
    shell:
        """
        samtools index 01split_bam/{wildcards.sample}/{wildcards.sample}_{wildcards.cluster_id}.bam
        """

def get_merge_bam_input(wildcards):
    samples = cluster_sample[wildcards.cluster_id]
    return expand("01split_bam/{sample}/{sample}_{{cluster_id}}.bam.bai", sample = samples)


rule merge_bam_per_cluster:
    input: get_merge_bam_input
    output: "01merged_bam/{cluster_id}.bam"
    params:
            bam = lambda wildcards, input: " ".join(input).replace(".bai", "")
    log: "00log/{cluster_id}.merge_bam.log"
    threads: 2
    shell:
        """
        samtools merge -@ 2 -r {output} {params.bam}
        """

它似乎起作用了。你知道吗