Constantin Ammar

The paper introduces QuantSelect, a deep learning method for fragment ion selection in DIA proteomics, improving quantification accuracy and differential expression sensitivity.

AlphaQuant improves proteoform regulation detection in bottom-up proteomics by using tree-based quantification and deep learning, revealing new patterns in mouse tissues.

AlphaDIA enables end-to-end transfer learning for feature-free proteomics, improving identification and quantification of post-translational modifications.

AlphaPept is an open-source framework for efficient processing of large MS-based proteomics datasets, achieving significant speed improvements using Python and machine learning.

The paper introduces scDVP, a spatial single-cell proteomics method, and finds that half of the murine hepatocyte proteome varies spatially within liver tissue.

The paper introduces MS-EmpiReS, a method to detect differential alternative splicing in proteomics data, identifying cancer-relevant splicing candidates and showing correlation with RNA sequencing.

This paper introduces a multiplex-DIA workflow that enhances single-cell proteome depth using dimethyl labeling and a reference channel, achieving reliable quantification of nearly 4,000 proteins per cell.

The paper introduces directLFQ, a label-free quantification method that scales linearly with sample numbers, enabling rapid analysis of large proteomics datasets.

The paper identifies proteins related to the bacterial envelope that improve starvation survival in E. coli by analyzing proteome adaptation across multiple conditions.

AlphaPeptDeep predicts peptide properties using deep learning, offering a modular framework that matches or exceeds existing tools for retention time, ion mobility, and fragment intensity prediction.

The paper develops a high-sensitivity mass spectrometry workflow to quantify single-cell proteome changes upon perturbation, revealing a stable core proteome despite transcriptomic variability.

EmpiReS improves detection of differentially expressed and alternatively spliced genes using fold change analysis with empirical error estimation.

MS-EmpiRe improves detection of differentially abundant proteins by using peptide-level noise distributions, doubling correct identifications at FDR cutoffs compared to existing tools.

The paper introduces a method to improve spectral library coverage in DIA proteomics by using multiple characteristic intensity patterns, achieving near-complete coverage without increasing false discovery rates.