篩選大型化合物庫以尋找藥物候選物是一項耗時的工作。此技能應用已確立的藥物化學規則,如 Lipinski 五規則、PAINS 過濾器和結構警示,自動識別有前景的化合物以供進一步研究。
تنزيل ZIP المهارة
رفع في Claude
اذهب إلى Settings → Capabilities → Skills → Upload skill
فعّل وابدأ الاستخدام
اختبرها
استخدام "medchem". Check if these SMILES pass drug-likeness rules: CC(=O)OC1=CC=CC=C1C(=O)O, c1ccccc1
النتيجة المتوقعة:
- Aspirin (CC(=O)OC1=CC=CC=C1C(=O)O): PASSES Rule of Five, PASSES Veber rules, no PAINS detected
- Benzene (c1ccccc1): FAILS Rule of Five (MW < 200), no alerts but poor drug candidate
استخدام "medchem". Filter this compound library for leadlike properties with max 400 complexity
النتيجة المتوقعة:
- Applied leadlike strict filter and Bertz complexity threshold
- Results: 127 of 500 compounds passed all filters
- Summary report saved to filtered_compounds_summary.txt
استخدام "medchem". Apply NIBR structural alerts and Lilly demerits to flag problematic molecules
النتيجة المتوقعة:
- NIBR filter: 89 compounds flagged for structural alerts
- Lilly demerits: 34 compounds exceed 100 demerits threshold
- Top alert patterns: anilines, Michael acceptors, reactive carbonyls
التدقيق الأمني
آمنPure medicinal chemistry library for molecule filtering. All 286 static findings are false positives: markdown code block delimiters flagged as shell commands, chemical terminology (filter, DES, complexity) mistaken for weak cryptography, and legitimate file I/O for chemical data processing. No network calls, credential access, or malicious behavior detected.
عوامل الخطر
📁 الوصول إلى نظام الملفات (2)
درجة الجودة
ماذا يمكنك بناءه
優先排序虛擬篩選結果
在實驗測試前篩選大型虛擬庫以識別類藥候選物。
支援先導物優化
在先導物優化期間應用更嚴格的標準以改善化合物可開發性。
庫的品質控制
從篩選集合中移除具有反應性基團或已知檢測干擾模式的化合物。
جرّب هذه الموجهات
Check if this SMILES [SMILES_STRING] passes the Rule of Five and has no PAINS alerts using medchem.
Filter this list of molecules for drug-likeness. Apply rule of five, rule of cns, and flag any with common structural alerts.
Apply leadlike strict criteria and Lilly demerits filter to these compounds. Report which ones pass all filters.
Run a comprehensive medchem analysis on this compound library: apply all relevant rules, check structural alerts, calculate complexity, and generate a summary report.
أفضل الممارسات
- 對大型庫使用平行化(n_jobs=-1)以加快篩選速度
- 結合多個過濾器(規則加上結構警示)進行全面品質控制
- 記錄哪些化合物被篩選以及原因,以確保可重現性
تجنب
- 盲目接受所有過濾結果而不進行領域知識審查
- 對所有標靶類別應用相同的過濾器而不進行調整
- 忽視已上市藥物經常違反標準規則的事實