跳转到内容

Contig ordering

作用:用 nucmer 把组装 contig 比对到参考组装,推导每个 query contig 的主导参考 contig、比对长度、比例与链向,用于排序与定向。 入口:05.1.nucmer.sh → 05.2.contigOrder.py 输入:Gifu_ref.fa(参考)与 Gifu_contig.fa(组装 contig),文件名硬编码并在工作目录解析。 输出:Gifu_scaffold.delta、Gifu_scaffold.filter.delta、Gifu_scaffold.filter.delta.coords、Gifu_scaffold.delta.coords、一张 mummerplot PNG;mummer_analysis_results/dominant_mapping.csv 与 .../query_contig_on_dominant_ref_colored.pdf(+ 蓝、- 红)。 运行:bash 05.1.nucmer.sh,然后 python 05.2.contigOrder.py 工具:MUMmer nucmer --prefix Gifu_scaffold、delta-filter -i 89 -l 1000 -1、show-coords -THrd、mummerplot --layout --png --large(需 gnuplot);Python 3 + pandas/matplotlib(read_csv(sep='\t', skiprows=5))。

English

Purpose: Align the assembled contigs to a reference with nucmer and derive each query contig’s dominant reference contig, aligned length, ratio and strand, for ordering and orientation. Entry point: 05.1.nucmer.sh → 05.2.contigOrder.py Inputs: Gifu_ref.fa (reference) and Gifu_contig.fa (contigs) — hardcoded filenames resolved in the working directory. Outputs: Gifu_scaffold.delta, Gifu_scaffold.filter.delta, Gifu_scaffold.filter.delta.coords, Gifu_scaffold.delta.coords, a mummerplot PNG; mummer_analysis_results/dominant_mapping.csv and .../query_contig_on_dominant_ref_colored.pdf (+ blue, - red). Run: bash 05.1.nucmer.sh, then python 05.2.contigOrder.py Tools: MUMmer nucmer --prefix Gifu_scaffold, delta-filter -i 89 -l 1000 -1, show-coords -THrd, mummerplot --layout --png --large (needs gnuplot); Python 3 with pandas and matplotlib (read_csv(sep='\t', skiprows=5)).

以下为该步骤目录中的实际文件(同步快照,与该目录内容一致)。

06.1.nucmer.sh · Shell · 9 行

在 GitHub 上查看此文件

终端窗口
nucmer --prefix Gifu_scaffold Gifu_ref.fa Gifu_contig.fa
delta-filter -i 89 -l 1000 -1 Gifu_scaffold.delta > Gifu_scaffold.filter.delta
show-coords -THrd Gifu_scaffold.filter.delta > Gifu_scaffold.filter.delta.coords
show-coords -THrd Gifu_scaffold.delta > Gifu_scaffold.delta.coords
mummerplot Gifu_scaffold.filter.delta -R Gifu_ref.fa -Q Gifu_contig.fa --layout --png --large
06.2.contigOrder.py · Python · 105 行

在 GitHub 上查看此文件

import pandas as pd
import matplotlib.pyplot as plt
import os
import seaborn as sns
# 输入输出设置
coords_path = "Gifu_scaffold.filter.delta.coords"
output_dir = "mummer_analysis_results"
os.makedirs(output_dir, exist_ok=True)
# 读取 coords 文件
df = pd.read_csv(coords_path, sep='\t', skiprows=5, header=None)
df.columns = [
"ref_start", "ref_end", "qry_start", "qry_end",
"ref_len", "qry_len", "percent_id", "ref_cov", "qry_cov",
"ref_contig", "qry_contig"
]
# 方向、长度等辅助列
df["strand"] = df.apply(lambda x: "+" if x["qry_start"] <= x["qry_end"] else "-", axis=1)
df["alignment_len"] = abs(df["qry_end"] - df["qry_start"]) + 1
df["ref_start_clean"] = df[["ref_start", "ref_end"]].min(axis=1)
df["ref_end_clean"] = df[["ref_start", "ref_end"]].max(axis=1)
# 汇总每条query对ref的比对长度与方向
summary_records = []
for qry, subdf in df.groupby("qry_contig"):
total_qry_len = subdf["qry_len"].iloc[0]
ref_stats = subdf.groupby("ref_contig").agg({
"alignment_len": "sum",
"strand": lambda x: list(x)
}).reset_index()
for _, row in ref_stats.iterrows():
ref = row["ref_contig"]
aligned_len = row["alignment_len"]
strands = row["strand"]
ratio = aligned_len / total_qry_len
plus = strands.count("+")
minus = strands.count("-")
dominant_strand = "+" if plus >= minus else "-"
summary_records.append([qry, ref, aligned_len, ratio, dominant_strand])
summary_df = pd.DataFrame(summary_records, columns=[
"Query_Contig", "Ref_Contig", "Aligned_Length", "Ratio", "Dominant_Strand"
])
# 找出每个 query 的 dominant ref
dominant_mapping = (
summary_df.groupby("Query_Contig")
.apply(lambda x: x.loc[x["Aligned_Length"].idxmax(), ["Query_Contig", "Ref_Contig", "Aligned_Length", "Dominant_Strand"]])
.reset_index(drop=True)
)
dominant_mapping["strand"] = dominant_mapping["Dominant_Strand"]
dominant_mapping.to_csv(os.path.join(output_dir, "dominant_mapping.csv"), index=False)
# 按 dominant ref 分组,绘制每个 query 的合并线段
plot_data = []
for _, row in dominant_mapping.iterrows():
qry = row["Query_Contig"]
ref = row["Ref_Contig"]
strand = row["strand"]
sub = df[(df["qry_contig"] == qry) & (df["ref_contig"] == ref)]
if sub.empty:
continue
start = sub["ref_start_clean"].min()
end = sub["ref_end_clean"].max()
plot_data.append([ref, qry, start, end, strand])
plot_df = pd.DataFrame(plot_data, columns=["ref_contig", "qry_contig", "start", "end", "strand"])
# 为每个 ref-query 分配独立的 y 坐标
plot_df["ref_qry"] = plot_df["ref_contig"] + "|" + plot_df["qry_contig"]
y_map = {}
y_val = 0
for ref in df["ref_contig"].unique():
ref_df = dominant_mapping[dominant_mapping["Ref_Contig"] == ref]
for qry in ref_df["Query_Contig"]:
key = f"{ref}|{qry}"
y_map[key] = y_val
y_val += 1
plot_df["y"] = plot_df["ref_qry"].map(y_map)
# ytick 按照 ref 显示
yticks = plot_df[["y", "ref_contig"]].drop_duplicates().sort_values("y")
# 画图
fig, ax = plt.subplots(figsize=(14, 0.5 * len(y_map) + 2))
for _, row in plot_df.iterrows():
color = "tab:blue" if row["strand"] == "+" else "tab:red"
ax.hlines(y=row["y"], xmin=row["start"], xmax=row["end"], color=color, linewidth=2)
ax.text((row["start"] + row["end"]) / 2, row["y"] + 0.3, row["qry_contig"], fontsize=6, ha="center", rotation=45)
ax.set_yticks(yticks["y"])
ax.set_yticklabels(yticks["ref_contig"])
ax.set_xlabel("Reference Genome Position")
ax.set_ylabel("Reference Contigs")
ax.set_title("Query Contig Mapping on Dominant Reference Contigs")
plt.tight_layout()
plt.savefig(os.path.join(output_dir, "query_contig_on_dominant_ref_colored.pdf"))
plt.close()
print("✔ dominant_mapping.csv 已保存")
print("✔ 主比对关系图 query_contig_on_dominant_ref_colored.pdf 已生成")