Agus Salim

bgnorm improves signal separation and marker positivity classification in multiplex spatial proteomics by addressing technical variation through background correction, normalization, and quality control.

This study evaluates deep learning methods for polygenic prediction, showing that MLP benefits most from larger sample sizes and one-hot encoding improves performance across methods.

SpaNorm improves spatial transcriptomics data normalisation by separating spatial library size effects from biological signals, enhancing domain and spatially variable gene identification.

This study shows that library size in spatial transcriptomics data confounds biological signals and recommends against normalizing it unless necessary.

MAPFX improves analysis of massively-parallel cytometry data by preprocessing, background correction, and imputing missing markers to enhance biological signal preservation and reduce technical variation.

This paper introduces RGStraP, a tool to infer genetic population structure from RNAseq data and shows it can control for population structure in gene expression analyses.

This study identifies major technical variations in microbiome data and evaluates a batch correction tool to minimize their impact on downstream analyses.

RUV-III-NB normalizes scRNA-seq data by removing unwanted variation from both cell embeddings and gene-level counts, improving differential expression and clustering results across multiple datasets.

scClassify improves cell type classification in scRNA-seq data by using ensemble learning, achieving more accurate and fine-grained results than original publications.

The paper develops a genomic risk score (metaGRS) for ischaemic stroke that outperforms existing risk factors in predicting stroke risk.

The paper identifies gene expression quantitative trait loci (eQTLs) in neonatal immune cells and links them to autoimmune and allergic disease risk through colocalisation and Mendelian randomisation.

DECENT improves differential expression analysis in scRNA-seq by modeling dropout, outperforming existing methods on simulated and real data without requiring spike-ins.

This study shows that ferrous-glutathione coupling promotes ferroptosis and frailty in C. elegans, and blocking ferroptosis extends lifespan and healthspan.