Trapiche

A Biome Classification Tool for Metagenomic Datasets

Trapiche is a hybrid method designed to classify metagenomic samples into specific biomes according to the GOLD Ecosystem Classification. It can utilize either text or taxonomy information independently to classify samples. However, integrating both text and taxonomy information yields the most accurate results.

Installation

To ensure a smooth installation process and avoid conflicts with other packages, we highly recommend installing Trapiche in a dedicated virtual environment. Follow these steps to install Trapiche:

pip install .

Basic usage

import glob

with open(file_with_text_description_of_project) as h: # 
    prj_desc = h.read()


from trapiche.llm_layer import lineages_from_text
from trapiche.deep_pred import predict_runs as deep_predict_runs

text_prediction = lineages_from_text(prj_desc)

taxonomy_files = glob.glob(f"{directory_with_taxonomy_annotations_files}/*mseq.txt")

result_dataframe = deep_predict_runs([[f] for f in taxonomy_files],return_full_preds=True,constrain=[prj_desc for _ in taxonomy_files])