Dilfuza Djamalova
Projects
Ongoing projects
AMR prediction in clinical isolates | July 2024 - Present
| ML-based trimming of nanopore sequences | July 2024 - Present |
PhD projects
Feature extraction pipeline | March 2021 - Present
- Built a snakemake pipeline to extract sequence-based features from bacterial genomes
- Pipeline takes the genome data of the bacterial strains as input and produces a comprehensive table with the computed features for each gene as an output
- Pipeline implements egnogg-mapper, esearch, FreeSASA, Foldseek, CodonW, and GOATOOLS along with custom python scripts
- Pipeline can be run with command
snakemake -p combine_features -j --cores 36 --use-conda, where-pspecifies the rule name,-jallows parallel execution of the non-connected rules,--coresallocates the required number of CPUs,--use-condaallows the python scripts or the required software to be executed in its conda environment.
Fig. Rulegraph of the feature extraction pipeline.
Bacterial fitness prediction | March 2021 - Present
- Implemented classifiers (XGBoost, Logistic Regression, Random Forest) using the Hydra Python framework.
- Pipeline architecture facilitates model training by allowing configuration files and data pre-processing steps (e.g. imputation of missing values) to be switched from the command line. It also allows new data processing or model training steps to be added without rewriting a significant amount of code.
Fig. Model training simplified scheme.
Code snippets used to generate plots | March 2021 - Present
- This repo contains Jupyter notebooks and Python scripts that are themselves small projects, e.g. to perform an incremental strain inclusion analysis, to assess a predictive power of sequence-based features, and to perform a cross-taxonomy prediction of bacterial essential genes at species, genera, family, and phylum levels.
- Also, Jupyter notebooks (1, 2, 3) contains example scripts to generate various plots.
Master’s project
Code snippets used in Master’s project | Sep 2018 - June 2020
- This repo contains python code snippets used in my Master’s project titled Re-classification of species and genera in family of Bacillaceae.
- Code snippets can be used to mainly collect bacterial genomes from NCBI Refseq/Genbank.
Coding practices
- I regularly try to update this repo with the solutions to ROSALIND tasks
- Repo contains Jupyter notebooks named after ROSALIND sections, e.g.
Python VillageorBioinformatics Stronghold.