196_基因组重排检测¶

# 运行Manta（肿瘤-正常配对模式）
configManta.py \
    --normalBam normal.bam \
    --tumorBam tumor.bam \
    --referenceFasta ref.fa \
    --runDir manta_output

python manta_output/runWorkflow.py -j 16 -g 32

# 输出
# manta_output/results/variants/
#   somaticSV.vcf.gz           # 体细胞SV
#   diploidSV.vcf.gz           # 胚系SV
#   candidateSmallIndels.vcf.gz # 小InDel候选
#   candidateSV.vcf.gz         # SV候选

# 查看体细胞SV
bcftools view manta_output/results/variants/somaticSV.vcf.gz | \
  bcftools query -f '%CHROM\t%POS\t%INFO/SVTYPE\t%INFO/SVLEN\t%FILTER\n' | head

# DELLY检测
delly call \
    -o delly_sv.bcf \
    -g ref.fa \
    tumor.bam normal.bam

# 过滤体细胞SV
echo -e "tumor\ttumor\nnormal\tcontrol" > samples.tsv
delly filter \
    -t DEL \
    -f somatic \
    -o delly_somatic_del.bcf \
    -s samples.tsv \
    delly_sv.bcf

# 检测所有类型
for svtype in DEL DUP INV BND; do
    delly filter -t $svtype -f somatic -o delly_somatic_${svtype}.bcf -s samples.tsv delly_sv.bcf
done

# 合并
bcftools concat delly_somatic_*.bcf | bcftools sort -o delly_somatic_all.vcf.gz

import pandas as pd
import matplotlib.pyplot as plt
import numpy as np

# 读取SV数据
sv = pd.read_csv("sv_calls.tsv", sep="\t")

# Chromothripsis的特征：
# 1. 一条染色体上大量SV（>10个断裂点）
# 2. 拷贝数在两个状态之间振荡
# 3. SV类型随机分布
# 4. 涉及片段大小相对均匀

# 每条染色体的SV密度
sv_per_chrom = sv.groupby('chrom1').size()
print("每条染色体的SV数:")
print(sv_per_chrom.sort_values(ascending=False))

# 检查高密度染色体是否符合chromothripsis特征
def detect_chromothripsis(sv_df, chrom, min_breakpoints=10):
    """简单的chromothripsis检测"""
    chrom_sv = sv_df[sv_df['chrom1'] == chrom].sort_values('pos1')

    if len(chrom_sv) < min_breakpoints:
        return False, {}

    # 检查SV类型分布（应该较均匀）
    type_counts = chrom_sv['svtype'].value_counts()
    type_entropy = -(type_counts / type_counts.sum() * np.log2(type_counts / type_counts.sum())).sum()

    # 检查断裂点间距分布（应该较均匀）
    distances = np.diff(chrom_sv['pos1'].values)
    distance_cv = np.std(distances) / np.mean(distances) if np.mean(distances) > 0 else float('inf')

    features = {
        'n_breakpoints': len(chrom_sv),
        'type_entropy': type_entropy,
        'distance_cv': distance_cv
    }

    # 高SV密度 + 类型混合 + 间距较均匀 → 可能是chromothripsis
    is_ct = len(chrom_sv) >= min_breakpoints and type_entropy > 1.0 and distance_cv < 2.0
    return is_ct, features

for chrom in sv_per_chrom[sv_per_chrom > 10].index:
    is_ct, features = detect_chromothripsis(sv, chrom)
    if is_ct:
        print(f"Potential chromothripsis on {chrom}: {features}")

library(ggplot2)
library(circlize)

# Circos图可视化重排
pdf("genome_rearrangements.pdf", width = 10, height = 10)
circos.initializeWithIdeogram(species = "hg38")

# 添加拷贝数轨道
cnv_data <- read.table("cnv_segments.bed", col.names = c("chr", "start", "end", "cn"))
circos.genomicTrack(cnv_data, panel.fun = function(region, value, ...) {
  circos.genomicLines(region, value, type = "segment", col = ifelse(value > 2, "red", "blue"))
})

# 添加SV连接线
sv_links <- read.table("sv_links.txt", col.names = c("chr1", "s1", "e1", "chr2", "s2", "e2", "type"))

# 按SV类型着色
type_colors <- c(DEL = "blue", DUP = "red", INV = "green", TRA = "orange")
for (i in 1:nrow(sv_links)) {
  circos.link(sv_links$chr1[i], c(sv_links$s1[i], sv_links$e1[i]),
              sv_links$chr2[i], c(sv_links$s2[i], sv_links$e2[i]),
              col = paste0(type_colors[sv_links$type[i]], "50"))
}

circos.clear()
dev.off()

# Manta体细胞SV检测
configManta.py --normalBam N.bam --tumorBam T.bam --referenceFasta ref.fa --runDir manta/
python manta/runWorkflow.py -j 16

# DELLY全类型检测
delly call -o sv.bcf -g ref.fa tumor.bam normal.bam

# GRIDSS高灵敏度检测
gridss -r ref.fa -o gridss.vcf.gz -t 16 --labels tumor,normal tumor.bam normal.bam

# SURVIVOR合并多工具结果
ls manta.vcf delly.vcf gridss.vcf > sv_list.txt
SURVIVOR merge sv_list.txt 1000 2 1 1 0 50 merged.vcf

# 统计SV类型
bcftools query -f '%INFO/SVTYPE\n' merged.vcf | sort | uniq -c