Advanced Adaptyv Platform

A scientific computing skill for working with the Adaptyv Bio platform — a cloud-based system for protein engineering, directed evolution experiments, and biological assay automation. Advanced Adaptyv Platform helps you design experiments, submit protein sequences for synthesis and testing, and analyze results from high-throughput screens.

When to Use This Skill

Choose Advanced Adaptyv Platform when:

• Designing directed evolution experiments for protein engineering
• Submitting protein variant libraries for high-throughput screening
• Analyzing assay results from Adaptyv Bio experiments
• Integrating Adaptyv API data into computational biology workflows

Consider alternatives when:

• You need general bioinformatics tools (use BioPython or similar)
• You're working with protein structure prediction only (use AlphaFold)
• You need sequence alignment or phylogenetics (use dedicated bioinformatics tools)

Quick Start

Copy
claude "Help me design a directed evolution experiment for GFP using Adaptyv"

Copy
# Example: Designing a protein variant library
import requests
import json

# Define the wild-type sequence
wild_type = "MSKGEELFTGVVPILVELDGDVNGHKFSVSGEGEGDAT..."

# Define mutation sites based on structural analysis
mutation_sites = [
    {"position": 65, "residues": ["S", "T", "A", "G"]},  # Chromophore
    {"position": 66, "residues": ["G", "A", "S"]},        # Chromophore
    {"position": 72, "residues": ["S", "T", "N", "Q"]},   # Folding
    {"position": 145, "residues": ["Y", "F", "W", "H"]},  # Fluorescence
    {"position": 203, "residues": ["T", "S", "A", "V"]},  # Stability
]

# Calculate library size
library_size = 1
for site in mutation_sites:
    library_size *= len(site["residues"])
print(f"Library size: {library_size} variants")

# Generate variant sequences
variants = generate_combinatorial_library(wild_type, mutation_sites)

Core Concepts

Directed Evolution Workflow

Step	Description	Output
Design	Select target protein and mutation strategy	Variant library definition
Synthesis	Gene synthesis of variant library	DNA constructs
Expression	Protein expression in host organism	Protein library
Screening	High-throughput functional assay	Activity measurements
Analysis	Statistical analysis of results	Improved variants
Iteration	Design next round based on results	Refined library

Library Design Strategies

Copy
# Site-saturation mutagenesis (NNK codon)
def nnk_library(sequence, positions):
    """Generate all 20 amino acids at specified positions"""
    amino_acids = "ACDEFGHIKLMNPQRSTVWY"
    library = []
    for pos in positions:
        for aa in amino_acids:
            variant = list(sequence)
            variant[pos] = aa
            library.append("".join(variant))
    return library

# Combinatorial library
def combinatorial_library(sequence, mutations):
    """Generate all combinations of specified mutations"""
    from itertools import product
    sites = [m["residues"] for m in mutations]
    positions = [m["position"] for m in mutations]

    variants = []
    for combo in product(*sites):
        variant = list(sequence)
        for pos, aa in zip(positions, combo):
            variant[pos] = aa
        variants.append("".join(variant))
    return variants

Result Analysis

Copy
import pandas as pd
import numpy as np

# Load screening results
results = pd.read_csv("screening_results.csv")

# Identify top performers
top_variants = results.nlargest(10, "activity_score")
print("Top 10 variants by activity:")
print(top_variants[["variant_id", "sequence", "activity_score", "stability"]])

# Epistasis analysis
def analyze_epistasis(results, positions):
    """Check for non-additive mutation effects"""
    singles = results[results["num_mutations"] == 1]
    doubles = results[results["num_mutations"] == 2]

    for _, double in doubles.iterrows():
        expected = sum(singles.loc[
            singles["mutation_pos"].isin(double["mutation_positions"])
        ]["delta_activity"])
        observed = double["delta_activity"]
        epistasis = observed - expected
        if abs(epistasis) > 0.5:
            print(f"Epistasis detected: {double['variant_id']} "
                  f"(expected={expected:.2f}, observed={observed:.2f})")

Configuration

Parameter	Description	Default
library_strategy	Saturation, combinatorial, or rational	combinatorial
max_library_size	Maximum variants to synthesize	1000
assay_type	Fluorescence, binding, activity, stability	Required
expression_host	E. coli, yeast, mammalian	e_coli
analysis_metrics	Metrics to report	[activity, stability]

Best Practices

1. Start with computational pre-screening. Before synthesizing a large library, use structure-based predictions (AlphaFold, Rosetta) to filter out variants likely to be non-functional. This reduces experimental costs and increases hit rates.

2. Include controls in every experiment. Always include the wild-type sequence and known positive/negative controls in your screening plate. Controls validate assay quality and provide a baseline for comparing variant performance.

3. Design libraries within platform capacity. Check the maximum library size the platform can handle in a single run. A 10,000-variant library is useless if the screening capacity is 1,000. Design libraries that fit within throughput constraints.

4. Analyze epistatic interactions systematically. When mutations at different positions interact non-additively, the best combinations can't be predicted from single-mutant data alone. Include enough double and triple mutants to detect epistasis.

5. Document experiment metadata thoroughly. Record expression conditions, assay parameters, plate layouts, and any deviations from protocol. Metadata that seems obvious today becomes critical for reproducing results six months later.

Common Issues

Library diversity lower than expected. Gene synthesis can introduce bias toward certain codons. Verify the actual library diversity with next-generation sequencing before screening. If diversity is low, the synthesis provider may need to adjust their process.

Screening results show high variance between replicates. Check for edge effects on plates (outer wells often behave differently), inconsistent cell density, or assay reagent degradation. Running 3+ replicates and removing outliers improves data quality.

Top variants don't reproduce in validation. Screening hits can be false positives from measurement noise. Always validate top candidates in triplicate with fresh preparations. If validation rates are low, the primary screen's signal-to-noise ratio needs improvement.