### 4-1. Quick Start

You can quickly get started by installing the official `openai` package in a Python environment and setting the API key.

```bash
# Install the official OpenAI Python SDK
pip install openai

# Set the API key environment variable
export OPENAI_API_KEY="sk-proj-YourOpenAIApiKeyHere..."
```

The following is an example of using Python to answer a simple bioinformatics question.

```python
from openai import OpenAI

# Initialize the API client (automatically loads the API key from the environment variable)
client = OpenAI()

response = client.chat.completions.create(
    model="gpt-4o",
    messages=[
        {"role": "system", "content": "You are a molecular biology expert. Please answer questions in an academic and detailed manner."},
        {"role": "user", "content": "Summarize the three most recent and representative guide RNA (gRNA) design techniques for reducing the off-target effects of CRISPR-Cas9."}
    ]
)

print(response.choices[0].message.content)
```

### 4-2. Detailed Installation

As of 2026, to implement the most widely used **Structured Outputs** in bioinformatics pipelines, this section provides installation and examples in conjunction with the `pydantic` library.

```bash
# Install the latest SDK and the Pydantic library for data validation
pip install --upgrade openai pydantic
```

```python
from pydantic import BaseModel, Field
from openai import OpenAI

# 1. Define the bio data structure to be extracted
class GeneAnnotation(BaseModel):
    gene_symbol: str = Field(description="Official gene symbol (e.g., BRCA1)")
    associated_disease: str = Field(description="Name of the associated disease")
    mutation_type: str = Field(description="Mutation type (e.g., Missense, Nonsense)")
    confidence: float = Field(description="Confidence score based on literature evidence (0.0 ~ 1.0)")

client = OpenAI()

# 2. Proceed with extracting structured data from the text
text_data = (
    "Recent studies have identified that the L858R missense mutation in the EGFR gene is one of the major causes of non-small cell lung cancer (NSCLC), and it has been reported to have a very high response rate to targeted therapies."
)

completion = client.beta.chat.completions.parse(
    model="gpt-4o",
    messages=[
        {"role": "system", "content": "Please extract gene annotation information accurately from the text."},
        {"role": "user", "content": text_data}
    ],
    response_format=GeneAnnotation,
)

# 3. Verify the results that have been safely parsed into an object
annotation = completion.choices[0].message.parsed
print(f"Gene: {annotation.gene_symbol}")
print(f"Disease: {annotation.associated_disease}")
print(f"Type: {annotation.mutation_type}")
print(f"Confidence: {annotation.confidence}")
```