interpret-nmr-spectrum
について
このスキルは、ClaudeがNMRスペクトル(1H、13C、DEPT、および2D実験)を解釈し、分子構造を決定することを可能にします。化学シフト、カップリングパターン、スペクトル積分値を分析して、構造フラグメントを提案します。NMRデータから有機化合物の構造を解明または確認する必要がある場合にご利用ください。
クイックインストール
Claude Code
推奨npx skills add pjt222/agent-almanac -a claude-code/plugin add https://github.com/pjt222/agent-almanacgit clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/interpret-nmr-spectrumこのコマンドをClaude Codeにコピー&ペーストしてスキルをインストールします
ドキュメント
Interpret NMR Spectrum
Read 1D and 2D NMR spectra. Assign peaks, determine coupling, propose molecular structural fragments consistent with all observed data.
When Use
- Determine structure of unknown organic compound from NMR data
- Confirm identity and purity of synthesized product
- Assign peaks in complex spectra with overlapping signals
- Correlate multiple NMR experiments (1H, 13C, DEPT, COSY, HSQC, HMBC) into unified structural picture
- Distinguish regioisomers, stereoisomers, conformational isomers
Inputs
- Required: NMR spectrum data (at minimum 1H spectrum with chemical shifts, multiplicities, integration)
- Required: Molecular formula or molecular weight (from mass spectrometry or elemental analysis)
- Optional: 13C and DEPT spectra (chemical shifts and multiplicities)
- Optional: 2D spectra (COSY, HSQC, HMBC, NOESY/ROESY correlation tables)
- Optional: Solvent and field strength used for acquisition
- Optional: Known structural constraints (e.g., reaction starting material, functional groups confirmed by IR)
Steps
Step 1: Assess Spectrum Type and Acquisition Parameters
Establish what data is available, its quality, before interpreting:
- Identify experiment types: Catalog which spectra available (1H, 13C, DEPT-135, DEPT-90, COSY, HSQC, HMBC, NOESY, ROESY, TOCSY). Note nucleus observed and dimensionality
- Record acquisition parameters: Spectrometer frequency (e.g., 400 MHz, 600 MHz), solvent, temperature, reference standard
- Identify solvent and reference peaks: Locate and exclude solvent signals using reference table:
| Solvent | 1H Residual (ppm) | 13C Signal (ppm) |
|---|---|---|
| CDCl3 | 7.26 | 77.16 |
| DMSO-d6 | 2.50 | 39.52 |
| D2O | 4.79 | -- |
| CD3OD | 3.31 | 49.00 |
| Acetone-d6 | 2.05 | 29.84, 206.26 |
| C6D6 | 7.16 | 128.06 |
- Assess spectral quality: Check baseline flatness, resolution of multiplets, signal-to-noise. Flag any artifacts (spinning sidebands, 13C satellites, solvent impurity peaks like H2O at ~1.56 ppm in CDCl3)
Got: Complete inventory of available experiments. Solvent/reference peaks excluded from analysis. Quality assessment.
If fail: Poor signal-to-noise or severe baseline distortion? Note limitation, proceed with caution. Flag any peaks not reliably distinguishable from noise.
Step 2: Analyze 1H Chemical Shifts
Assign each 1H signal to chemical environment using characteristic shift ranges:
- Tabulate all signals: For each peak, record chemical shift (ppm), multiplicity, coupling constant(s) J (Hz), relative integration
- Classify by chemical shift region:
| Range (ppm) | Environment | Examples |
|---|---|---|
| 0.0--0.5 | Shielded (cyclopropane, M-H) | Cyclopropyl H, metal hydrides |
| 0.5--2.0 | Alkyl (CH3, CH2, CH) | Saturated aliphatic chains |
| 2.0--4.5 | Alpha to heteroatom/unsaturation | -OCH3, -NCH2, allylic, benzylic |
| 4.5--6.5 | Vinyl / olefinic | =CH-, =CH2 |
| 6.5--8.5 | Aromatic | ArH |
| 9.0--10.0 | Aldehyde | -CHO |
| 10.0--12.0 | Carboxylic acid | -COOH |
| 0.5--5.0 (broad, exchangeable) | OH, NH | Alcohols, amines, amides |
- Count hydrogens: Use integration ratios relative to molecular formula to assign number of protons per signal. Normalize to simplest whole-number ratio
- Note exchangeable protons: Signals disappearing on D2O shake (OH, NH, COOH) are exchangeable. Record presence and approximate shift
Got: Table of all 1H signals with shift, multiplicity, J-values, integration (number of H), preliminary environment assignment.
If fail: Integration ratios not summing to expected total protons? Check for overlapping signals, broad peaks hidden in baseline, or incorrect molecular formula.
Step 3: Determine Coupling Patterns and J-Values
Extract connectivity info from splitting patterns:
- Identify multiplicities: Assign each signal as singlet (s), doublet (d), triplet (t), quartet (q), doublet of doublets (dd), etc. For complex multiplets (m), estimate number of coupling partners
- Measure coupling constants: Extract J-values in Hz. Match reciprocal couplings (if H_A couples to H_B with J = 7.2 Hz, H_B must show same J to H_A)
- Classify J-values by type:
| J Range (Hz) | Coupling Type |
|---|---|
| 0--3 | Geminal (2J) or long-range (4J, 5J) |
| 6--8 | Vicinal aliphatic (3J) |
| 8--10 | Vicinal with restricted rotation |
| 10--17 | Vicinal olefinic cis (6--12) or trans (12--18) |
| 0--3 | Aromatic meta |
| 6--9 | Aromatic ortho |
- Map coupling networks: Group mutually coupled protons into spin systems. Each spin system = connected fragment of molecule
- Assess roof effect: In AB-type patterns, inner lines of doublets more intense than outer → chemical shift proximity
Got: All coupling constants measured and matched reciprocally. Spin systems identified. Coupling types classified.
If fail: Multiplets too complex for first-order rules? Note higher-order pattern. Overlapping signals or strongly coupled nuclei (delta-nu/J < 10) make non-first-order patterns needing simulation.
Step 4: Analyze 13C and DEPT Data
Determine carbon types and count from 13C experiments:
- Count distinct carbon signals: Compare number of 13C peaks vs molecular formula. Fewer peaks than expected = molecular symmetry
- Classify by chemical shift:
| Range (ppm) | Carbon Type | Examples |
|---|---|---|
| 0--50 | sp3 Alkyl | CH3, CH2, CH, quaternary C |
| 50--100 | Alpha to O or N | -OCH3, -OCH2, anomeric C |
| 100--150 | Aromatic / vinyl | =CH-, ArC |
| 150--170 | Heteroaromatic / enol / imine | C=N, C-O aromatic |
| 170--185 | Carboxyl / ester / amide | -COOH, -COOR, -CONR2 |
| 185--220 | Aldehyde / ketone | -CHO, >C=O |
- Apply DEPT editing: Use DEPT-135 (CH and CH3 up, CH2 down, quaternary absent) and DEPT-90 (CH only) to determine number of attached hydrogens per carbon
- Calculate degree of unsaturation: DBE = (2C + 2 + N - H - X) / 2. Compare with count of pi bonds and rings implied by spectrum
Got: Every 13C signal classified by type (CH3, CH2, CH, C) and chemical environment. Degree of unsaturation calculated, consistent with observed functional groups.
If fail: DEPT data unavailable? Infer hydrogen attachment from HSQC correlations (Step 5). Carbon count does not match molecular formula? Check for coincident signals or quaternary carbons hidden in noise.
Step 5: Correlate 2D NMR Data
Build connectivity using two-dimensional experiments:
- COSY (1H-1H correlation): Identify which protons are 2-3 bonds apart. Map cross-peaks to confirm and extend spin systems from Step 3
- HSQC (1H-13C one-bond): Assign each proton to directly bonded carbon. Links 1H and 13C assignments unambiguously
- HMBC (1H-13C long-range): Identify 2-3 bond H-C correlations. HMBC critical for connecting fragments across quaternary carbons, heteroatoms, carbonyl groups lacking direct H-C bonds
- NOESY/ROESY (through-space): Identify protons spatially close (< 5 Angstroms) regardless of bonding. Use for stereochemical assignment and conformational analysis
- Build fragment connectivity: Use HMBC correlations to connect spin systems from COSY into larger fragments. Each HMBC cross-peak = 2-3 bond path from H to C
Got: Connectivity map linking all spin systems into coherent molecular framework, with stereochemical info from NOE data where available.
If fail: 2D data incomplete or ambiguous? Note which connections are tentative. Multiple structural proposals may be necessary. Prioritize HMBC correlations for fragment assembly — bridges gaps COSY cannot.
Step 6: Propose and Validate Structure
Assemble fragments into complete structural proposal:
- Assemble fragments: Connect structural fragments from Steps 2-5 using HMBC correlations and degree-of-unsaturation constraints
- Check molecular formula: Verify proposed structure matches molecular formula exactly (atom count, degree of unsaturation)
- Back-predict chemical shifts: For proposed structure, predict expected 1H and 13C chemical shifts. Compare with observed. Deviations > 0.3 ppm (1H) or > 5 ppm (13C) warrant re-examination
- Verify all correlations: Confirm every observed COSY, HSQC, HMBC correlation explained by proposed structure. Unexplained cross-peaks = error or impurity
- Consider alternatives: Multiple structures fit the data? List distinguishing experiments or correlations that would resolve ambiguity
- Assign stereochemistry: Use NOE data, J-value analysis (Karplus relationship for dihedral angles), known conformational preferences to assign relative and, where possible, absolute stereochemistry
Got: Single best-fit structural proposal with all NMR data accounted for, or ranked list of candidates with plan to distinguish them.
If fail: No single structure accounts for all data? Check for: mixture of compounds (extra peaks with non-integer integration ratios), dynamic processes (broad peaks from conformational exchange), paramagnetic impurities (anomalous broadening). Re-examine molecular formula if multiple structures remain equally viable.
Checks
- All solvent and reference peaks identified and excluded from interpretation
- Every 1H signal assigned chemical shift region, multiplicity, J-value, integration
- Coupling constants reciprocal (matched between coupling partners)
- 13C signals classified by DEPT multiplicity and chemical shift region
- Degree of unsaturation calculated and consistent with proposed structure
- 2D correlations (COSY, HSQC, HMBC) all explained by structural proposal
- Proposed structure matches molecular formula exactly
- Back-predicted chemical shifts agree with observed values within tolerance
- Stereochemistry addressed using NOE and/or J-value analysis where applicable
Pitfalls
- Ignoring solvent peaks: Common solvents make signals overlapping analyte peaks. Always identify and exclude solvent residuals, water, grease peaks before interpretation.
- Forcing first-order analysis on second-order patterns: Strongly coupled nuclei (small chemical shift difference relative to J) make distorted multiplets not interpretable with simple n+1 rules. Recognize roof effects and non-binomial intensity patterns as indicators.
- Overlooking exchangeable protons: OH and NH signals may be broad, shifted by concentration/temperature, or absent in protic solvents. D2O shake experiment clarifies which signals exchangeable.
- Assuming all 13C peaks visible: Quaternary carbons have long relaxation times and low intensity. May be absent from short-acquisition spectra. HMBC correlations often only way to detect them.
- Misinterpret HMBC artifacts: HMBC spectra can show one-bond artifacts (misassigned as long-range correlations) and weak four-bond correlations. Cross-check with HSQC to filter out one-bond leakthrough.
- Neglect symmetry: Observed number of 13C peaks fewer than molecular formula predicts? Molecule likely has symmetry element. Account for this before proposing structure.
See Also
interpret-ir-spectrum— identify functional groups to constrain NMR-based structure proposalsinterpret-mass-spectrum— determine molecular formula and fragmentation for cross-validationinterpret-uv-vis-spectrum— characterize chromophores and conjugation extentinterpret-raman-spectrum— complementary vibrational data for symmetric modesplan-spectroscopic-analysis— select and sequence spectroscopic techniques before data acquisition
GitHub リポジトリ
関連スキル
llamaguard
その他LlamaGuardは、暴力やヘイトスピーチなど6つの安全性カテゴリーにおいて、LLMの入力と出力をモデレートするMetaの70-80億パラメータモデルです。94〜95%の精度を提供し、vLLM、Hugging Face、Amazon SageMakerを使用してデプロイ可能です。このスキルを使用して、AIアプリケーションにコンテンツフィルタリングと安全策を簡単に統合できます。
cost-optimization
その他このClaudeスキルは、リソースの適正サイジング、タグ付け戦略、支出分析を通じて、開発者がクラウドコストを最適化することを支援します。AWS、Azure、GCPにわたるクラウド支出の削減とコストガバナンスの実施のためのフレームワークを提供します。インフラコストの分析、リソースの適正サイジング、または予算制約への対応が必要な際にご利用ください。
quantizing-models-bitsandbytes
その他このスキルは、bitsandbytesを使用してLLMを8ビットまたは4ビット精度に量子化し、精度の低下を最小限に抑えつつ50〜75%のメモリ削減を実現します。限られたGPUメモリでより大規模なモデルを実行したり、推論を高速化するのに理想的で、INT8、NF4、FP4などのフォーマットをサポートしています。HuggingFace Transformersと統合され、QLoRAトレーニングや8ビットオプティマイザーを可能にします。
dispatching-parallel-agents
その他このClaudeスキルは、複数のエージェントを配備し、3つ以上の独立した問題を並行して調査・修正します。共有状態や依存関係がなく解決可能な、無関係な障害が発生するシナリオ向けに設計されています。中核となる機能は並列問題解決であり、効率を最大化するために独立した問題領域ごとに1つのエージェントを割り当てます。
