MSstatsShiny: A GUI for Versatile, Scalable, and Reproducible Statistical Analyses of Quantitative Proteomic Experiments

Kohler, Devon; Kaza, Maanasa; Pasi, Cristina; Huang, Ting; Staniak, Mateusz; Mohandas, Dhaval; Sabidó, Eduard; Choi, Meena; Vitek, Olga

doi:10.1021/acs.jproteome.2c00603

Cited by 15 publications

(15 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For more complex designs we recommend MSstats, which can model a variety of experimental designs in a statistically rigorous manner. 55 We found that imputation can identify new quantitative peptides (Figure 5). As modern proteomics techniques increase the number of identifications, it is important to remember that not all of the detected peptides are quantitative.…”

Section: Discussionmentioning

confidence: 93%

Evaluating Proteomics Imputation Methods with Improved Criteria

Harris,

Fondrie,

et al. 2023

J. Proteome Res.

View full text Add to dashboard Cite

Quantitative measurements produced by tandem mass spectrometry proteomics experiments typically contain a large proportion of missing values. Missing values hinder reproducibility, reduce statistical power, and make it difficult to compare across samples or experiments. Although many methods exist for imputing missing values, in practice, the most commonly used methods are among the worst performing. Furthermore, previous benchmarking studies have focused on relatively simple measurements of error such as the mean-squared error between imputed and held-out values. Here we evaluate the performance of commonly used imputation methods using three practical, “downstream-centric” criteria. These criteria measure the ability to identify differentially expressed peptides, generate new quantitative peptides, and improve the peptide lower limit of quantification. Our evaluation comprises several experiment types and acquisition strategies, including data-dependent and data-independent acquisition. We find that imputation does not necessarily improve the ability to identify differentially expressed peptides but that it can identify new quantitative peptides and improve the peptide lower limit of quantification. We find that MissForest is generally the best performing method per our downstream-centric criteria. We also argue that existing imputation methods do not properly account for the variance of peptide quantifications and highlight the need for methods that do.

show abstract

Section: Discussionmentioning

confidence: 93%

Evaluating Proteomics Imputation Methods with Improved Criteria

Harris,

Fondrie,

et al. 2023

J. Proteome Res.

View full text Add to dashboard Cite

show abstract

“…Therefore, it would be great to assess further if machine learning models can be trained on lower-level data as peptides are the most sensible unit and imputation on protein groups level performs one form of implicit imputation at the peptide level 8 . One could assess if imputed features on lower-level data can be reaggregated to protein groups, e.g using ideas from Sticker and coworkers 43 or MSStats 44 .…”

Section: Discussionmentioning

confidence: 99%

“…We did not include non-general imputation methods as e.g. provided by MSstats 44 or without reusable software 72 . We used KNN interpolation of replicates based on the HeLa cell line measurements being repeated over time.…”

Section: Other Imputation Approachesmentioning

confidence: 99%

Imputation of label-free quantitative mass spectrometry-based proteomics data using self-supervised deep learning

Webel

Niu

Nielsen

et al. 2023

Preprint

View full text Add to dashboard Cite

Imputation techniques provide means to replace missing measurements with a value and are used in almost all downstream analysis of mass spectrometry (MS) based proteomics data using label-free quantification (LFQ). Some methods only impute assuming the limit of detection (LOD) was not passed and therefore impute missing values with too low or too high intensities, potentially leading to biased results in downstream statistical analysis. Here we test how self supervised deep learning models can impute missing values in the context of LFQ at different levels: precursors, aggregated peptides or protein groups. We demonstrate how collaborative filtering, denoising autoencoders, and variational autoencoders can be used to reconstruct missing values and can make more relevant features available for downstream analysis compared to current approaches. Additionally, we show that deep learning approaches can model data in its entirety for imputation and offer an approach for controlled evaluation of imputation approaches. We applied our method, proteomics imputation modeling mass spectrometry (PIMMS), to an alcohol-related liver disease (ALD) cohort with blood plasma proteomics data available for 358 individuals. We identified 49 additional proteins (+52.7%) that are significantly differentially abundant across disease stages compared to traditional methods and found that some of these were predictive of ALD progression in machine learning models. We, therefore, suggest the use of deep learning approaches for imputing missing values in MS-based proteomics and provide workflows for these.

show abstract

“…The intensities were compared using LFQ (label-free quantification) values [80] with Perseus (version 2.0.11 [81]). Contaminants and reverse sequences were removed from the matrix and LFQ intensities were log2-transformed using Mstats [82]. Before ttest and visualization using a volcano plot, the missing values were replaced by imputation with the normal distribution for each column separately (default settings).…”

Section: Protein Co-ip and Analyses By Lc-ms/msmentioning

confidence: 99%

Interactors and effects of overexpressing YlxR/RpnM, a conserved RNA binding protein in cyanobacteria

Hemm,

Miucci,

Riediger

et al. 2024

Preprint

View full text Add to dashboard Cite

Throughout the tree of life RNA-binding proteins play important roles in the regulation of gene expression and RNA metabolism, but they are only poorly characterized in cyanobacteria. Here, we analyzed the predicted RNA-binding protein Ssr1238 from the model cyanobacteriumSynechocystissp. PCC 6803. Ssr1238 is encoded in a syntenic region within thenusA-ssr1238-infBoperon, similar to genes encoding such proteins in gram-positive bacteria. Overexpression of Ssr1238 for 24 h led to slightly higher levels of RNase P RNA, 44 tRNAs, and several stress-related mRNAs. Co-immunoprecipitation of proteins followed by MS analysis and sequencing of UV crosslinked, co-immunoprecipitated RNA samples identified potential interaction partners of Ssr1238. The most highly enriched transcript was RNase P RNA, and RnpA, the protein component of RNase P, was among the most highly enriched proteins, consistent with recent findings that YlxR proteins can influence the enzymatic activity of RNase P. A second highly enriched transcript derived from the 3’ region of genessl3177, which encodes a rare lipoprotein homolog, a central enzyme in cell wall remodeling during cell division. The data also showed a strong connection to the RNA maturation and modification system indicated by co-precipitation of RNA methyltransferase Sll1967, the A-adding tRNA nucleotidyltransferase Sll1253, and queuine tRNA-ribosyltransferase Slr0713, as well as riboendonuclease E and enolase. Surprisingly, cyanophycin synthetase and urease were highly enriched as well. Therefore, Ssr1238 specifically binds to two different transcripts and appears to participate in the coordination of RNA maturation, aspects of nitrogen metabolism and translation, and cell division. Our results are consistent with recent findings that theB. subtilisYlxR protein functions as an RNase P modulator (RnpM), extend its proposed role to the phylum cyanobacteria, and suggest additional functionalities.

show abstract

MSstatsShiny: A GUI for Versatile, Scalable, and Reproducible Statistical Analyses of Quantitative Proteomic Experiments

Cited by 15 publications

References 31 publications

Evaluating Proteomics Imputation Methods with Improved Criteria

Evaluating Proteomics Imputation Methods with Improved Criteria

Imputation of label-free quantitative mass spectrometry-based proteomics data using self-supervised deep learning

Interactors and effects of overexpressing YlxR/RpnM, a conserved RNA binding protein in cyanobacteria

Contact Info

Product

Resources

About