为了满足 NCBI 的严格要求，并体现所有序列均为环状（circular），我们需要在 FASTA 头部添加 [topology=circular]。同时，我为你设计了一套符合国际命名规范的质粒名称（以 p 开头，结合亚型缩写和编号，去除了连字符以防止 NCBI 解析报错）。

1. 质粒命名设计 (Plasmid Names)

基于你提供的映射表，质粒将被命名为 p{Subtype}_{Number}：

• WT 亚型: pWT_1, pWT_2 …
• O 亚型: pO_1, pO_2 …
• T 亚型: pT_1, pT_2 …
• O-T (O_Trans): pOT_1, pOT_2 …
• T-T (WT_Trans): pTT_1, pTT_2 … (去除了连字符以防NCBI解析错误)
• S2-O (S2_Dark): pS2O_1, pS2O_2 …
• S2-T (S2_Light): pS2T_1, pS2T_2 …

2. 自动化修复脚本

这个 Python 脚本会自动完成以下工作：

1. 识别最长的 contig 标记为 [location=chromosome]。
2. 将较短的 contig 标记为 [plasmid-name=pXXX_1]。
3. 为所有序列（染色体和质粒）添加 [topology=circular]。
4. 自动备份原文件为 .original_backup。

在你的 FASTA 文件所在目录下（trycycler-medaka_polished_genomes），创建并运行以下脚本：

nano fix_ncbi_circular.py

复制以下代码，保存并退出 (Ctrl+O, Enter, Ctrl+X)：

#!/usr/bin/env python3
import os
import shutil

# 映射文件名到亚型名称
FILE_MAPPING = {
    "WT-trycycler-medaka_polished_genome.fa": "WT",
    "O-trycycler-medaka_polished_genome.fa": "O",
    "T-trycycler-medaka_polished_genome.fa": "T",
    "WT_Trans-trycycler-medaka_polished_genome.fa": "T-T",
    "O_Trans-trycycler-medaka_polished_genome.fa": "O-T",
    "S2_Dark-trycycler-medaka_polished_genome.fa": "S2-O",
    "S2_Light-trycycler-medaka_polished_genome.fa": "S2-T"
}

ORGANISM = "Acinetobacter baumannii"

def process_fasta(filepath, subtype):
    sequences = []
    with open(filepath, 'r') as f:
        header = None
        seq_parts = []
        for line in f:
            line = line.strip()
            if not line:
                continue
            if line.startswith('>'):
                if header is not None:
                    sequences.append((header, ''.join(seq_parts)))
                # 提取纯净的ID，去掉原来的多余描述
                header = line[1:].split()[0]
                seq_parts = []
            else:
                seq_parts.append(line)
        if header is not None:
            sequences.append((header, ''.join(seq_parts)))

    if not sequences:
        print(f"⚠️ 文件为空: {filepath}")
        return

    # 按长度降序排序（最长的是染色体）
    sequences.sort(key=lambda x: len(x[1]), reverse=True)

    # 备份原文件 (只备份一次)
    backup_path = filepath + ".original_backup"
    if not os.path.exists(backup_path):
        shutil.copy(filepath, backup_path)

    # 重写文件
    plasmid_idx = 1
    with open(filepath, 'w') as out_f:
        for i, (hdr, seq) in enumerate(sequences):
            if i == 0:
                # 最长的序列 -> 染色体
                new_hdr = f">{hdr} [organism={ORGANISM}] [location=chromosome] [topology=circular]"
            else:
                # 较短的序列 -> 质粒
                # 将 T-T 转换为 TT 以避免 ID 中的特殊字符问题
                safe_subtype = subtype.replace("-", "") 
                p_name = f"p{safe_subtype}_{plasmid_idx}"
                new_hdr = f">{hdr} [organism={ORGANISM}] [plasmid-name={p_name}] [topology=circular]"
                plasmid_idx += 1

            out_f.write(new_hdr + '\n')
            # NCBI 要求序列按 80 个字符一行折叠
            for j in range(0, len(seq), 80):
                out_f.write(seq[j:j+80] + '\n')

    print(f"✅ 成功处理: {filepath} (1 个染色体, {plasmid_idx - 1} 个质粒)")

print("==========================================")
print("🧬 开始格式化 FASTA 文件以符合 NCBI 规范")
print("==========================================")

for filename, subtype in FILE_MAPPING.items():
    if os.path.exists(filename):
        process_fasta(filename, subtype)
    else:
        print(f"❌ 未找到文件: {filename}")

print("\n🎉 所有文件格式化完成！现在可以重新上传至 NCBI。")

3. 运行脚本

python3 fix_ncbi_circular.py

4. 脚本运行后的 FASTA 头部示例

脚本会将你的 FASTA 头部修改为完美的 NCBI 格式，例如对于 WT 样本：

>cluster_001_consensus [organism=Acinetobacter baumannii] [location=chromosome] [topology=circular]
ATCGATCG... (80 chars per line)
>cluster_002_consensus [organism=Acinetobacter baumannii] [plasmid-name=pWT_1] [topology=circular]
ATCGATCG... (80 chars per line)
>cluster_005_consensus [organism=Acinetobacter baumannii] [plasmid-name=pWT_2] [topology=circular]
ATCGATCG... (80 chars per line)

5. 重新上传

修改完成后，直接在 NCBI 提交页面的 Files 选项卡中重新上传这 7 个修改后的 .fa 文件。 因为现在每个文件都明确声明了 chromosome、plasmid-name 以及 topology=circular，NCBI 的验证器（Validator）将会直接放行，不会再报 FATAL 错误。

---

---

Yes. In the script, all contigs are first sorted by length in descending order:

sequences.sort(key=lambda x: len(x[1]), reverse=True)

Then the naming proceeds in that order:

• Index 0 (longest, ~3.9 Mb) → [location=chromosome]
• Plasmid _1 → the longest plasmid
• Plasmid _2 → the second‑longest plasmid
• …and so on.

Example (WT): if the plasmids are 100 kb, 60 kb and 20 kb, they will be named:

• pWT_1 = 100 kb
• pWT_2 = 60 kb
• pWT_3 = 20 kb

Note that the output FASTA is also rewritten in this sorted order (chromosome first, then plasmids from largest to smallest), which is a common convention for finished bacterial genomes.

If you would rather have the smallest plasmid = _1 (ascending), just change the plasmid assignment to count from the end:

n_plasmids = len(sequences) - 1
...
else:
    p_name = f"p{safe_subtype}_{n_plasmids - plasmid_idx + 1}"
    plasmid_idx += 1

But as written, yes — _1 is always the largest plasmid.