This repository includes a python implemenation of XGMix, a gradient boosting tree-based local-ancestry inference (ancestry deconvolution) method.
XGMIX.py can be used in two ways:
- training a model from scratch using provided training data or
- loading a pre-trained XGMix model (see Pre-Trained Models below)
In both cases the models are used to infer local ancestry for provided query data.
The dependencies are listed in requirements.txt. Assuming pip is already installed, they can be installed via
$ pip install -r requirements.txt
When using the program for training a model, BCFtools must be installed and available in the PATH environment setting.
XGMIX.py loads and uses pre-trained XGMix models to predict the ancestry for a given <query_file> and chromosome number.
To execute the program with a pre-trained model run:
$ python3 XGMIX.py <query_file> <genetic_map_file> <output_basename> <chr_nr> <path_to_model>
where
- <query_file> is a .vcf or .vcf.gz file containing the query haplotypes which are to be analyzed (see example in the /demo_data folder)
- <genetic_map_file> is the genetic map file (see example in the /demo_data folder)
- <output_basename>.msp.tsv. is where the predictions are written (see details in Output below and an example in the /demo_data folder)
- <chr_nr> is the chromosome number
- <path_to_model> is a path to the model used for predictions (see Pre-trained Models below)
XGMix.py loads data from the <reference_file>
To execute the program when training a model run:
$ python3 XGMIX.py <query_file> <genetic_map_file> <output_basename> <chr_nr> <reference_file> <sample_map_file>
where the first 4 arguments are described above in the pre-trained setting and
- <reference_file> is a .vcf or .vcf.gz file containing the reference haplotypes (in any order)
- <sample_map_file> is a sample map file matching reference samples to their respective reference populations
The program uses these two files as input to rfmix's simulation algorithm to create training data for the model.
XGmix output probabilities might not reflect the true confidence / accuracy of the predictions. By setting to True calibration when training a new model, Isotonic Regression is used to match the predicted probabilities to calibrated probabilities. For example, in a calibrated model, predictions with a probability 80% will be correct 80% of the time.
More advanced configuration settings can be found in config.py. They include general settings, simulation settings and model settings. More details are given in the file itself.
The first line is a comment line, that specifies the order and encoding of populations: eg: #Sub_population order/code: golden_retriever=0 labrador_retriever=1 poodle poodle_small=2
The second line specifies the column names, and every following line marks the genome position.
The first 6 columns specify
- the chromosome
- interval of genetic marker's physical position in basepair units (one column represents the starting point and one the end point)
- interval of genetic position in centiMorgans (one column represents the starting point and one the end point)
- number of <query_file> SNP positions that are included in interval
The remaining columns give the predicted reference panel population for the given interval. A genotype has two haplotypes, so the number of predictions for a genotype is 2*(number of genotypes) and therefore the total number of columns in the file is 6 + 2*(number of genotypes)
When training a model, the resulting model will be stored in ./models. That way it can be re-used for analyzing another dataset. The model's estimated accuracy is logged along with a confusion matrix which is stored in ./models/analysis. The program simulates training data and stores in ./generated_data. To automatically remove the created data when training is done, set rm_simulated_data to True in config.py. Note that in some cases, the simulated data can be re-used for training with similar settings. In those cases, not removing the data and then setting run_simulation to False will re-use the previously simulated data which can save a lot of time and compuation.
Pre-trained models are available for download from XGMix-models.
When making predictions, the input to the model is an intersection of the pre-trained model SNP positions and the SNP positions from the <query_file>. That means that the set of positions that's only in the original training input is encoded as missing and the set of positions only in the <query_file> is discarded. When the script is executed, it will log the intersection-ratio as the performance will depend on how much of the original positions are missing. When the intersection is low, we recommend using a model trained with high percentage of missing data.
The models are trained on hg build 37 references from the following biogeographic regions: Subsaharan African (AFR), African Hunter and Gatherer (AHG), East Asian (EAS), European (EUR), Native American (NAT), Oceanian (OCE), South Asian (SAS), and West Asian (WAS) and labels and predicts them as 0, 1, .., 7 respectively.
When using this software, please cite: Kumar, A., Montserrat, D.M., Bustamante, C. and Ioannidis, A., "XGMix: Local-Ancestry Inference With Stacked XGBoost," International Conference on Learning Representations Workshops (ICLR, 2020, Workshop AI4AH).
https://www.biorxiv.org/content/10.1101/2020.04.21.053876v1
@article{kumar2020xgmix,
title={XGMix: Local-Ancestry Inference With Stacked XGBoost},
author={Kumar, Arvind and Montserrat, Daniel Mas and Bustamante, Carlos and Ioannidis, Alexander},
journal={International Conference of Learning Representations Workshops, AI4AH},
year={2020}
}
You can also include its companion paper: Montserrat, D.M., Kumar, A., Bustamante, C. and Ioannidis, A., "Addressing Ancestry Disparities in Genomic Medicine: A Geographic-aware Algorithm," International Conference on Learning Representations Workshops (ICLR, 2020, Workshop AI4CC).
https://arxiv.org/pdf/2004.12053.pdf
@article{montserrat2020addressing,
title={Addressing Ancestry Disparities in Genomic Medicine: A Geographic-aware Algorithm},
author={Montserrat, Daniel Mas and Kumar, Arvind and Bustamante, Carlos and Ioannidis, Alexander},
journal={International Conference of Learning Representations Workshops, AI4CC},
year={2020}
}