• Solid mathematical background: algebra, probability & statistics
• R, Python, SAS
• Linux
• Data structures & algorithms
• Fundamentals of Neural Networks & Machine Learning
• Bioinformatics technologies (Next Generation Sequencing, Microarray Genotyping, Comparative hybridization, ChiP-Seq, Hi-C, PCR)
• Pipelines creation with Snakemake
• Molecular biology fundamentals
• Latex

My experience includes:

• Participation in the project devoted to studying genomic integrity of patients with cancer 

(since 09.2020 - 04.2021)

I was responsible for selection of suitable tools for structural variants detection and following pipelines creation with Snakemake (Delly, CNVnator, LUMPY, MELT, GRIDSS). I worked with alignments & variant calls through samtools and bcftools. My colleagues & I worked on development of wavelet-based genomic instability metrics. I tried different coverage normalizations to see their impact on the metrics proposed. Also I analyzed Genome In a Bottle papers & metrics behaviour on their gold standard data. In one of my latest tasks I needed to analyse coverage wave-forms for patterns in certain types of genomic rearrangements for db creation. 

• Teaching at St.-Petersburg Electrotechnical University “LETI” (since 09.2019 - till now)

Up to now I’ve been teaching 3 types of seminars for 1-3 year students. Namely, linear algebra, probability & statistics. 1-semester statistics course covers: rigorous point estimation (ML, REML), UMVUE theory & construction, hypothesis testing, quadratic forms distribution, LRT, Wald test, UMP tests, confidence ellipsoids, LM, GLM, GLMM, ANOVA, contingency tables analysis, multiple testing issues, basics of epidemiological studies designs.

• Development of a new mathematically rigorous method to identify genomic regions associated with phenotypes in my Bachelor’s final project. The tool is implemented in R (09.2020 - till now)

I had to estimate the correlation matrix of a multivariate normal vector constructed from partitioned chi-squared statistics from a sequence of contingency tables. First of all, I had to find such an orthogonal transformation so that the chi-squared statistic with (n-1)(m-1) df is a sum of squares of (n-1)(m-1) iid normal variables. For the needs of genetics 2x2 & 3x2 tables have been considered. Point estimate of the correlation matrix has been obtained. A method for controlling for PSD-property of the matrix has been proposed.

• Analysis of genotyping data in alcohol dependence study (09.2020-05.2021)

Under supervision of a biologist & a mathematician worked with genotyping data of a St.-Petersburg cohort of alcohol dependent subjects. Tasks started from QC & preprocessing, followed by GWAS-analysis and ended with the results analysis and comparison with known associations from publicly available GWAS databases (DOI:10.1016/j.euroneuro.2021.08.262)

• Neural networks construction & training

During the machine learning course I used Python language and Keras library for neural networks creation & training. I worked with dense, convolutional & recurrent NN for analysis of numeric and text data. Tasks included classification, regression and text generation.