AI based alignment free (agnostic) taxonomic classification. Public deliverable 9.2

The purpose of this work was to demonstrate the usage of artificial
intelligence/machine learning for metagenomics analysis. More specifically,
we aimed to implement an AI-based analysis toolkit for human
papillomavirus sequence identification that will be further used to try to
predict cancer development based on HPV status and other
metatranscriptomes present in the specimen.
Datasets used for the study comprised the whole cancer genome atlas
(TCGA, https://www.cancer.gov/tcga) database, a well-known and
validated public database and part of the Swedish cervical screening cohort
as well as the Finnish HPV vaccination cohort.
In this deliverable, we describe several metagenomics pipelines such as the
HPV-Meta and the microbial metagenome pipeline used to produce the input
data for the Machine Learning model. We provide the description of the
process of creating the Machine Learning software to analyse the data, the
challenges, the analysis and results, as well as the future plan for the
improvement of the HEAP AI-based software toolkit for metagenomics
analysis.