Julián Candia

The paper develops a proteomic signature, MitoAge, that predicts age-related mitochondrial health based on circulating factors in serum.

This paper introduces SomaModules, a pathway enrichment approach for SomaScan data, showing improved enrichment compared to original gene sets in Alzheimer's Disease and mitochondrial pathway analyses.

This study evaluates how in vitro hemolysis and freeze-thaw cycles affect protein abundance measurements in SomaScan and Olink assays, identifying affected analytes and proposing quality control measures.

This study examines changes in motor unit conduction velocity after unilateral lower limb suspension and active recovery, finding that ion channel gene expression, particularly K+ transporters, correlates with these changes.

The paper describes bioinformatic methods for normalization, quality control, and variability assessment of SomaScan proteomics data.

Across Baltimore Longitudinal Study of Aging replicates, fully normalized SomaScan 11K has median inter-plate coefficient of variation 5% (7K: 4.5%), intra-plate 3%, 7K inter-batch 8%.

This study evaluates GSEA performance on RNA-seq data using TCGA and curated pathways, finding the classic permutation method offers better sensitivity-specificity tradeoffs than complex alternatives.

The paper identifies age-associated DNA methylation sites in immune cells and links them to transcription factor motifs and histone modifications, suggesting adaptive responses to oxygen fluctuations in aging.

Using inter-plate duplicates from 102 donors, a self-contained normalization of the 7k SomaScan plasma assay yields median CV 4.5% across 7,288 human-protein SOMAmers.

This paper uses an explainable machine learning framework to predict and explain cognitive differences from task-based fMRI data, finding that ML models outperform traditional methods in prediction and provide consistent explanations.

mutSigMapper provides a method to map mutational spectra to signatures using shot-noise modeling, offering a non-parametric statistical assessment of significance.

eNetXplorer is an R package that evaluates elastic net models for generalized linear models using cross-validation and provides tools for feature selection and significance testing.

RefCell analyzes image-based HTS data by identifying typical cells in reference states to assess deviations from typical behavior in perturbations or samples.

This paper introduces a methanol fixation method for preserving PBMCs for scRNA-seq, showing it maintains RNA integrity and transcriptional profiles comparable to live cells.