Skip to content
View pkamand1's full-sized avatar

Block or report pkamand1

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
pkamand1/README.md

Patricia Kamanda

I am a data scientist working across biomedical signal processing, multi-omics, and health informatics. My work pairs statistical modeling and machine learning with several years as an applications scientist in the sequencing industry, where I built and supported analysis pipelines for research and clinical laboratories.

What I work on

  • Physiological signal processing and deep learning for ECG in Python, including the validation of wearable device measurements against clinical reference signals (NeuroKit2, HeartPy, statsmodels, PyTorch)
  • Transcriptomic, proteomic, and metabolomic analysis, from bulk and single-cell RNA-seq to RPPA and LC-MS data
  • Reproducible sequencing pipelines in Nextflow with Docker and Singularity
  • Clinical and longitudinal data, with mixed models, survival analysis, SQL, and SAS
  • Clinical operations and trial data: scheduling and billing analytics, CDISC SDTM, edit checks, and risk-based monitoring
  • Analytics engineering and decision science: layered SQL warehouses, dbt, query tuning, forecasting, and optimization
  • ClinInfoJobs.com, a job board I built and run for health informatics roles that do not require a clinical license, automated end to end with GitHub Actions

Selected projects

Signal processing and machine learning

Project What it shows
ecg-heartbeat-classification-pytorch A 1D residual network in PyTorch against classical baselines, evaluated on unseen patients

Clinical and health data

Project What it shows
ambulatory-clinic-operations-analytics A SQL star schema for Power BI, no-show prediction, a difference-in-differences evaluation of text reminders, provider productivity, charge lag, and coding review
clinical-trial-data-management Edit checks and query listings, protocol deviations, CDISC SDTM mapping with XPT output and conformance checks, and risk-based site monitoring
pbc-longitudinal-survival Linear mixed models, Cox and time-dependent Cox models, and competing risks in R
sas-sql-longitudinal-cohort Cohort construction, data quality checks, and medication adherence in MySQL and SAS

Analytics engineering and optimization

Project What it shows
game-telemetry-sql-analytics A layered MySQL warehouse with stored procedures, views, partitioning, and measured query tuning; the same models in dbt with tests; retention, stability, and bot detection
inventory-fulfillment-optimization An explainable demand forecast, deterministic and stochastic linear programs, Monte Carlo policy evaluation, and replenishment simulation

Multi-omics

Project What it shows
nf-viral-consensus A Nextflow pipeline for viral whole-genome sequencing, from QC to phylogeny
nf-metagenomic-classification A Nextflow pipeline for quality control, host read removal, and taxonomic classification of short and long reads
scrnaseq-pbmc-interferon Single-cell RNA-seq in Python with Scanpy, Harmony integration, and pseudobulk PyDESeq2
scrnaseq-pseudobulk-benchmark Why cell-level tests flood multi-donor studies with false positives, and how pseudobulk avoids it, on simulated data with known truth
rnaseq-deseq2-paired-design A paired-design bulk RNA-seq analysis with DESeq2
bulk-rnaseq-de-benchmark DESeq2, edgeR, limma-voom, and simpler tests compared on simulated counts: false discovery control, power, and outliers
tcga-brca-rppa-proteomics limma subtype differences, clustering, and survival models on TCGA proteomics
melanoma-pdx-rppa-treatment-analysis Protein and phosphoprotein responses of melanoma xenografts to BRAF and MEK inhibition, dasatinib, and their combination
urine-metabolomics-age-bmi-sex Untargeted urinary metabolomics with univariate testing and OPLS-DA
ckd-exercise-serum-metabolomics Targeted serum metabolomics of chronic kidney disease with and without voluntary exercise
metabolomics-drift-imputation-benchmark QC-based drift correction and missing-value imputation for LC-MS, compared on simulated studies
culex-pipiens-sex-biased-expression-reanalysis A reanalysis of published RNA-seq expression data (Martynova, Kamanda & Sim, 2022)

Tools

Python, R, SQL (MySQL, DuckDB, dbt), SAS, Nextflow, Docker, Singularity, Scanpy, DESeq2, edgeR, limma, scikit-learn, statsmodels, PuLP, PyTorch, Power BI, Tableau, Plotly Dash, and GitHub Actions.

Find me

clininfojobs.com · ORCID

Pinned Loading

  1. scrnaseq-pbmc-interferon scrnaseq-pbmc-interferon Public

    Single-cell RNA-seq of IFN-beta-stimulated PBMCs in Python: Scanpy QC and clustering, Harmony integration, marker annotation, and pseudobulk PyDESeq2

    Jupyter Notebook

  2. ecg-heartbeat-classification-pytorch ecg-heartbeat-classification-pytorch Public

    Deep learning for ECG heartbeat classification in PyTorch: a 1D residual network with RR-interval features on the MIT-BIH Arrhythmia Database, evaluated on unseen patients against classical baselines

    Python

  3. ambulatory-clinic-operations-analytics ambulatory-clinic-operations-analytics Public

    Medical group operations analytics: SQL star schema for Power BI, no-show prediction, difference-in-differences evaluation of reminders, provider productivity, charge lag, and coding review

    HTML

  4. clinical-trial-data-management clinical-trial-data-management Public

    Clinical trial data management: edit checks and query listings, protocol deviations, CDISC SDTM mapping with XPT output and conformance checks, and risk-based site monitoring

    Python

  5. game-telemetry-sql-analytics game-telemetry-sql-analytics Public

    Layered game telemetry warehouse in MySQL and dbt: stored procedures, views, partitioning, query tuning, data quality checks, bot detection, and retention analysis on simulated data

    Python

  6. nf-viral-consensus nf-viral-consensus Public

    Nextflow pipeline for viral WGS: QC, Bowtie2 alignment, variant calling, masked consensus genomes, coverage QC, and IQ-TREE phylogeny

    Nextflow