Supported Models

AlignAIR ships with pre-trained models for human B-cell and T-cell receptor analysis. Each model is optimized for specific chain types and comes ready to use in the Docker container.

Quick Reference

IGH Heavy Chain

--chain-type=heavy

/app/pretrained_models/IGH_S5F_576

IGL/IGK Light Chain

--chain-type=light

/app/pretrained_models/IGL_S5F_576

TCRB Beta Chain

--chain-type=tcrb

/app/pretrained_models/TCRB_UNIFORM_576

Model Details

IGH Heavy Chain

Immunoglobulin Heavy Chain model trained on S5F mutation patterns

Key Features

V/D/J segmentation

Allele calling

Mutation prediction

Productivity assessment

Usage Example:

python app.py run \
  --model-checkpoint=/app/pretrained_models/IGH_S5F_576 \
  --chain-type=heavy \
  --sequences=/data/input/sequences.csv \
  --save-path=/data/output

Model Metadata

Checkpoint Path

/app/pretrained_models/IGH_S5F_576

Chain Type

heavy

Species

Human

Reference Set

OGRDB V8 extended

Last Updated

February 2025

Input Size

576 nucleotides

IGL/IGK Light Chain

Immunoglobulin Lambda Light Chain model with enhanced V/J prediction

Key Features

V/J segmentation

Allele calling

Mutation prediction

Productivity assessment

Usage Example:

python app.py run \
  --model-checkpoint=/app/pretrained_models/IGL_S5F_576 \
  --chain-type=light \
  --sequences=/data/input/sequences.csv \
  --save-path=/data/output

Model Metadata

Checkpoint Path

/app/pretrained_models/IGL_S5F_576

Chain Type

light

Species

Human

Reference Set

OGRDB V2 & V3 extended

Last Updated

March 2025

Input Size

576 nucleotides

TCRB Beta Chain

T Cell Receptor Beta Chain model optimized for TCR repertoire analysis

Key Features

V/D/J segmentation

Allele calling

Productivity assessment

Usage Example:

python app.py run \
  --model-checkpoint=/app/pretrained_models/TCRB_UNIFORM_576 \
  --chain-type=tcrb \
  --sequences=/data/input/sequences.csv \
  --save-path=/data/output

Model Metadata

Checkpoint Path

/app/pretrained_models/TCRB_UNIFORM_576

Chain Type

tcrb

Species

Human

Reference Set

IMGT 2022

Last Updated

July 2025

Input Size

576 nucleotides

Model Comparison

Model	Chain Type	Segments	Reference	Use Case
IGH Heavy Chain	`heavy`	V, D, J	OGRDB V8 extended	B-cell heavy chain analysis
IGL/IGK Light Chain	`light`	V, J	OGRDB V2 & V3 extended	B-cell lambda/kappa light chain analysis
TCRB Beta Chain	`tcrb`	V, D, J	IMGT 2020	T-cell receptor beta chain analysis

All Models Included

All models are pre-installed and ready to use when you pull the AlignAIR Docker container. No additional downloads or setup required.

docker pull thomask90/alignair:latest