@nahisaho/satori 0.13.0 → 0.14.0

package/src/.github/skills/scientific-phylogenetics/SKILL.md

@@ -0,0 +1,297 @@
+---
+name: scientific-phylogenetics
+description: |
+  系統解析スキル。ete3/ETE Toolkit による系統樹構築・可視化、
+  scikit-bio 系統的多様性、配列アライメントベース進化解析、
+  分子時計・分岐年代推定、祖先配列再構成パイプライン。
+---
+
+# Scientific Phylogenetics
+
+ETE Toolkit / scikit-bio を中心とした
+分子系統解析・進化生物学パイプラインを提供する。
+
+## When to Use
+
+- 分子系統樹を構築・可視化するとき (NJ/ML/ベイズ法)
+- 多重配列アライメントから系統推定するとき
+- 分岐年代推定 (分子時計) を行うとき
+- 系統的多様性 (PD: Phylogenetic Diversity) を計算するとき
+- 祖先配列再構成を行うとき
+- 系統比較法 (PGLS 等) で形質進化を解析するとき
+
+---
+
+## Quick Start
+
+## 1. ETE Toolkit 系統樹構築
+
+```python
+from ete3 import Tree, TreeStyle, NodeStyle, faces, AttrFace
+import subprocess
+import tempfile
+
+
+def build_phylogenetic_tree(sequences_fasta, method="fasttree", model="GTR"):
+    """
+    配列アライメントから系統樹構築。
+
+    Parameters:
+        sequences_fasta: str — FASTA ファイルパス (アライン済み)
+        method: str — "fasttree", "raxml", "iqtree"
+        model: str — 進化モデル ("GTR", "JTT", "WAG", "LG")
+
+    K-Dense: etetoolkit — Phylogenetics toolkit
+    """
+    commands = {
+        "fasttree": ["fasttree", "-gtr", "-nt", sequences_fasta],
+        "raxml": [
+            "raxmlHPC", "-s", sequences_fasta, "-n", "tree",
+            "-m", f"GTRGAMMA", "-p", "12345",
+        ],
+        "iqtree": [
+            "iqtree2", "-s", sequences_fasta,
+            "-m", model, "-bb", "1000", "--prefix", "iqtree_out",
+        ],
+    }
+
+    cmd = commands.get(method, commands["fasttree"])
+    result = subprocess.run(cmd, capture_output=True, text=True)
+
+    if method == "fasttree":
+        newick = result.stdout
+    elif method == "iqtree":
+        with open("iqtree_out.treefile", "r") as f:
+            newick = f.read()
+    else:
+        newick = result.stdout
+
+    tree = Tree(newick)
+    print(f"Phylogenetic tree ({method}, {model}): "
+          f"{len(tree)} leaves, {len(list(tree.traverse()))} total nodes")
+    return tree
+
+
+def visualize_tree(tree, output_file="phylogenetic_tree.png",
+                   layout="rectangular", show_support=True):
+    """
+    ETE3 系統樹可視化。
+
+    Parameters:
+        tree: ete3.Tree — 系統樹オブジェクト
+        output_file: str — 出力画像パス
+        layout: str — "rectangular", "circular"
+        show_support: bool — ブートストラップ値を表示
+    """
+    ts = TreeStyle()
+    ts.mode = "c" if layout == "circular" else "r"
+    ts.show_leaf_name = True
+    ts.show_branch_length = True
+    ts.show_branch_support = show_support
+    ts.branch_vertical_margin = 10
+
+    # Node styling
+    for node in tree.traverse():
+        nstyle = NodeStyle()
+        if node.is_leaf():
+            nstyle["fgcolor"] = "#2196F3"
+            nstyle["size"] = 8
+        else:
+            nstyle["fgcolor"] = "#E91E63"
+            nstyle["size"] = 5
+            if show_support and node.support >= 0.9:
+                nstyle["fgcolor"] = "#4CAF50"
+        node.set_style(nstyle)
+
+    tree.render(output_file, tree_style=ts, w=800, units="px")
+    print(f"Tree rendered: {output_file} ({layout} layout)")
+    return output_file
+```
+
+## 2. 多重配列アライメント
+
+```python
+from Bio import AlignIO, SeqIO
+from Bio.Align.Applications import MafftCommandline, MuscleCommandline
+
+
+def run_multiple_alignment(input_fasta, method="mafft", output_fasta=None):
+    """
+    多重配列アライメント。
+
+    Parameters:
+        input_fasta: str — 入力 FASTA パス
+        method: str — "mafft", "muscle", "clustalw"
+        output_fasta: str — 出力パス
+    """
+    if output_fasta is None:
+        output_fasta = input_fasta.replace(".fasta", f"_aligned_{method}.fasta")
+
+    if method == "mafft":
+        cmd = f"mafft --auto {input_fasta} > {output_fasta}"
+    elif method == "muscle":
+        cmd = f"muscle -in {input_fasta} -out {output_fasta}"
+    else:
+        cmd = f"clustalw2 -INFILE={input_fasta} -OUTFILE={output_fasta}"
+
+    subprocess.run(cmd, shell=True, check=True)
+
+    alignment = AlignIO.read(output_fasta, "fasta")
+    print(f"Alignment ({method}): {len(alignment)} sequences, "
+          f"{alignment.get_alignment_length()} positions")
+    return alignment
+```
+
+## 3. 系統的多様性 (Phylogenetic Diversity)
+
+```python
+import skbio
+from skbio import TreeNode
+from skbio.diversity import alpha_diversity, beta_diversity
+import numpy as np
+
+
+def calculate_phylogenetic_diversity(newick_string, sample_otus):
+    """
+    系統的多様性 (Faith's PD, UniFrac) 計算。
+
+    Parameters:
+        newick_string: str — Newick 形式系統樹
+        sample_otus: dict — {sample_id: {otu_id: abundance}}
+
+    K-Dense: scikit-bio — PD & UniFrac
+    """
+    tree = TreeNode.read([newick_string])
+
+    # Prepare OTU table
+    all_otus = sorted(set(
+        otu for otus in sample_otus.values() for otu in otus
+    ))
+    sample_names = list(sample_otus.keys())
+    otu_table = np.zeros((len(sample_names), len(all_otus)))
+    for i, sample in enumerate(sample_names):
+        for j, otu in enumerate(all_otus):
+            otu_table[i, j] = sample_otus[sample].get(otu, 0)
+
+    # Faith's PD (alpha diversity)
+    pd_values = alpha_diversity("faith_pd", otu_table, ids=sample_names, tree=tree,
+                                otu_ids=all_otus)
+    print(f"Faith's PD: mean={pd_values.mean():.3f}, "
+          f"range=[{pd_values.min():.3f}, {pd_values.max():.3f}]")
+
+    # Weighted UniFrac (beta diversity)
+    unifrac_dm = beta_diversity("weighted_unifrac", otu_table,
+                                ids=sample_names, tree=tree, otu_ids=all_otus)
+    print(f"Weighted UniFrac: mean distance = "
+          f"{unifrac_dm.condensed_form().mean():.4f}")
+
+    return {"faith_pd": pd_values, "unifrac": unifrac_dm}
+```
+
+## 4. 分子時計・分岐年代推定
+
+```python
+def estimate_divergence_times(tree, calibrations, rate_model="strict"):
+    """
+    分子時計による分岐年代推定。
+
+    Parameters:
+        tree: ete3.Tree — 系統樹
+        calibrations: dict — {(taxon1, taxon2): (min_age, max_age)}
+            e.g., {("human", "mouse"): (85, 95)}  # MYA
+        rate_model: str — "strict" or "relaxed"
+    """
+    # Branch length to relative time conversion
+    total_length = max(tree.get_distance(leaf) for leaf in tree.get_leaves())
+
+    # Apply calibration
+    for (t1, t2), (min_age, max_age) in calibrations.items():
+        node1 = tree.search_nodes(name=t1)
+        node2 = tree.search_nodes(name=t2)
+        if node1 and node2:
+            ancestor = tree.get_common_ancestor(node1[0], node2[0])
+            dist = tree.get_distance(ancestor)
+            calibration_age = (min_age + max_age) / 2
+            rate = dist / calibration_age if calibration_age > 0 else 1
+            print(f"Calibration {t1}-{t2}: {calibration_age} MYA, rate={rate:.6f}")
+
+    # Estimate ages for all internal nodes
+    node_ages = {}
+    for node in tree.traverse("postorder"):
+        if not node.is_leaf():
+            dist = tree.get_distance(node)
+            # Simple proportional dating
+            estimated_age = (dist / total_length) * max(
+                (min_age + max_age) / 2
+                for (min_age, max_age) in calibrations.values()
+            )
+            node_ages[node.name or f"node_{id(node)}"] = estimated_age
+
+    return node_ages
+```
+
+## 5. 祖先配列再構成
+
+```python
+def ancestral_sequence_reconstruction(alignment_file, tree_file, model="JTT"):
+    """
+    最尤法による祖先配列再構成。
+
+    Parameters:
+        alignment_file: str — アライメントファイルパス
+        tree_file: str — 系統樹ファイルパス (Newick)
+        model: str — アミノ酸置換モデル
+    """
+    # Using IQ-TREE for ASR
+    cmd = [
+        "iqtree2", "-s", alignment_file, "-te", tree_file,
+        "-m", model, "-asr", "--prefix", "asr_output",
+    ]
+    result = subprocess.run(cmd, capture_output=True, text=True)
+
+    if result.returncode == 0:
+        # Parse ancestral sequences
+        asr_file = "asr_output.state"
+        ancestral_seqs = {}
+        if os.path.exists(asr_file):
+            import csv
+            with open(asr_file) as f:
+                reader = csv.reader(f, delimiter="\t")
+                for row in reader:
+                    if row and not row[0].startswith("#"):
+                        node = row[0]
+                        site = row[1]
+                        state = row[2]
+                        if node not in ancestral_seqs:
+                            ancestral_seqs[node] = []
+                        ancestral_seqs[node].append(state)
+
+        print(f"ASR ({model}): {len(ancestral_seqs)} ancestral nodes reconstructed")
+        return ancestral_seqs
+    else:
+        print(f"ASR failed: {result.stderr[:200]}")
+        return None
+```
+
+---
+
+## パイプライン出力
+
+| 出力ファイル | 説明 | 連携先スキル |
+|---|---|---|
+| `results/phylogenetic_tree.nwk` | Newick 系統樹 | → infectious-disease, microbiome |
+| `figures/phylogenetic_tree.png` | 系統樹可視化 | → publication-figures, presentation |
+| `results/divergence_times.json` | 分岐年代推定 | → population-genetics, environmental-ecology |
+| `results/ancestral_sequences.fasta` | 祖先配列 | → protein-design, sequence-analysis |
+| `results/phylo_diversity.json` | 系統的多様性 | → microbiome-metagenomics |
+
+## パイプライン統合
+
+```
+sequence-analysis ──→ phylogenetics ──→ infectious-disease
+  (アライメント)      (系統樹構築)     (病原体系統解析)
+                           │
+                           ├──→ microbiome-metagenomics (UniFrac)
+                           ├──→ population-genetics (分岐推定)
+                           └──→ environmental-ecology (系統的多様性)
+```