Bayesian Confidence Intervals for Multiplexed Proteomics Integrate Ion-statistics with Peptide Quantification Concordance

A data analysis framework for combining multiple batches increases the power of isobaric proteomics experiments

We present a framework for the analysis of multiplexed mass spectrometry proteomics data that reduces estimation error when combining multiple isobaric batches. Variations in the number and quality of observations have long complicated the analysis of isobaric proteomics data. Here we show that the power to detect statistical associations is substantially improved by utilizing models that directly account for known sources of variation in the number and quality of observations that occur across batches.In a multibatch benchmarking experiment, our open-source software (msTrawler) increases the power to detect changes, especially in the range of less than twofold changes, while simultaneously increasing quantitative proteome coverage by utilizing more low-signal observations. Further analyses of previously published multiplexed datasets of 4 and 23 batches highlight both increased power and the ability to navigate complex missing data patterns without relying on unverifiable imputations or discarding reliable measurements.

The sulfotransferase XB5850668.L is required to apportion embryonic ectodermal domains

BACKGROUND

Members of the sulfotransferase superfamily (SULT) influence the activity of a wide range of hormones, neurotransmitters, metabolites and xenobiotics. However, their roles in developmental processes are not well characterized even though they are expressed during embryogenesis. We previously found in a microarray screen that Six1 up-regulates LOC100037047, which encodes XB5850668.L, an uncharacterized sulfotransferase.

RESULTS

Since Six1 is required for patterning the embryonic ectoderm into its neural plate, neural crest, preplacodal and epidermal domains, we used loss- and gain-of function assays to characterize the role of XB5850668.L during this process. Knockdown of endogenous XB5850668.L resulted in the reduction of epidermal, neural crest, cranial placode and otic vesicle gene expression domains, concomitant with neural plate expansion. Increased levels had minimal effects, but infrequently expanded neural plate and neural crest gene domains, and infrequently reduced cranial placode and otic vesicle gene domains. Mutation of two key amino acids in the sulfotransferase catalytic domain required for PAPS binding and enzymatic activity tended to reduce the effects of overexpressing the wild-type protein.

CONCLUSIONS

Our analyses indicates that XB5850668.L is a member of the SULT2 family that plays important roles in patterning the embryonic ectoderm. Some aspects of its influence likely depend on sulfotransferase activity.

The trouble with triples: Examining the impact of measurement error in mediation analysis

Mediation analysis is used in genetic mapping studies to identify candidate gene mediators of quantitative trait loci (QTL). We consider genetic mediation analysis of triplets-sets of three variables consisting of a target trait, the genotype at a QTL for the target trait, and a candidate mediator that is the abundance of a transcript or protein whose coding gene co-locates with the QTL. We show that, in the presence of measurement error, mediation analysis can infer partial mediation even in the absence of a causal relationship between the candidate mediator and the target. We describe a measurement error model and a corresponding latent variable model with estimable parameters that are combinations of the causal effects and measurement errors across all three variables. The relative magnitudes of the latent variable correlations determine whether or not mediation analysis will tend to infer the correct causal relationship in large samples. We examine case studies that illustrate the common failure modes of genetic mediation analysis and demonstrate how to evaluate the effects of measurement error. While genetic mediation analysis is a powerful tool for identifying candidate genes, we recommend caution when interpreting mediation analysis findings.

Challenges and Opportunities for Bayesian Statistics in Proteomics

Proteomics is a data-rich science with complex experimental designs and an intricate measurement process. To obtain insights from the large data sets produced, statistical methods, including machine learning, are routinely applied. For a quantity of interest, many of these approaches only produce a point estimate, such as a mean, leaving little room for more nuanced interpretations. By contrast, Bayesian statistics allows quantification of uncertainty through the use of probability distributions. These probability distributions enable scientists to ask complex questions of their proteomics data. Bayesian statistics also offers a modular framework for data analysis by making dependencies between data and parameters explicit. Hence, specifying complex hierarchies of parameter dependencies is straightforward in the Bayesian framework. This allows us to use a statistical methodology which equals, rather than neglects, the sophistication of experimental design and instrumentation present in proteomics. Here, we review Bayesian methods applied to proteomics, demonstrating their potential power, alongside the challenges posed by adopting this new statistical framework. To illustrate our review, we give a walk-through of the development of a Bayesian model for dynamic organic orthogonal phase-separation (OOPS) data.

The Intriguing Landscape of Single-Cell Protein Analysis

Profiling protein expression at single-cell resolution is essential for fundamental biological research (such as cell differentiation and tumor microenvironmental examination) and clinical precision medicine where only a limited number of primary cells are permitted. With the recent advances in engineering, chemistry, and biology, single-cell protein analysis methods are developed rapidly, which enable high-throughput and multiplexed protein measurements in thousands of individual cells. In combination with single cell RNA sequencing and mass spectrometry, single-cell multi-omics analysis can simultaneously measure multiple modalities including mRNAs, proteins, and metabolites in single cells, and obtain a more comprehensive exploration of cellular signaling processes, such as DNA modifications, chromatin accessibility, protein abundance, and gene perturbation. Here, the recent progress and applications of single-cell protein analysis technologies in the last decade are summarized. Current limitations, challenges, and possible future directions in this field are also discussed.

Quantitative Accuracy and Precision in Multiplexed Single-Cell Proteomics

Single-cell proteomics workflows have considerably improved in sensitivity and reproducibility to characterize as-yet unknown biological phenomena. With the emergence of multiplexed single-cell proteomics, studies increasingly present single-cell measurements in conjunction with an abundant congruent carrier to improve the precursor selection and enhance identifications. While these extreme carrier spikes are often >100× more abundant than the investigated samples, the total ion current undoubtably increases but the quantitative accuracy possibly is affected. We here focus on narrowly titrated carrier spikes (i.e., <20×) and assess their elimination for a comparable sensitivity with superior accuracy. We find that subtle changes in the carrier ratio can severely impact the measurement variability and describe alternative multiplexing strategies to evaluate data quality. Lastly, we demonstrate elevated replicate overlap while preserving acquisition throughput at an improved quantitative accuracy with DIA-TMT and discuss optimized experimental designs for multiplexed proteomics of trace samples. This comprehensive benchmarking gives an overview of currently available techniques and guides the conceptualization of the optimal single-cell proteomics experiment.

TMTpro Complementary Ion Quantification Increases Plexing and Sensitivity for Accurate Multiplexed Proteomics at the MS2 Level

Multiplexed proteomics is a powerful tool to assay cell states in health and disease, but accurate quantification of relative protein changes is impaired by interference from co-isolated peptides. Interference can be reduced by using MS3-based quantification, but this reduces sensitivity and requires specialized instrumentation. An alternative approach is quantification by complementary ions, the balancer group-peptide conjugates, which allows accurate and precise multiplexed quantification at the MS2 level and is compatible with most proteomics instruments. However, complementary ions of the popular TMT-tag form inefficiently and multiplexing is limited to five channels. Here, we evaluate and optimize complementary ion quantification for the recently released TMTpro-tag, which increases complementary ion plexing capacity to eight channels (TMTproC). Furthermore, the beneficial fragmentation properties of TMTpro increase sensitivity for TMTproC, resulting in ∼65% more proteins quantified compared to TMTpro-MS3 and ∼18% more when compared to real-time-search TMTpro-MS3 (RTS-SPS-MS3). TMTproC quantification is more accurate than TMTpro-MS2 and even superior to RTS-SPS-MS3. We provide the software for quantifying TMTproC data as an executable that is compatible with the MaxQuant analysis pipeline. Thus, TMTproC advances multiplexed proteomics data quality and widens access to accurate multiplexed proteomics beyond laboratories with MS3-capable instrumentation.

Precise Temporal Regulation of Post-transcriptional Repressors Is Required for an Orderly Drosophila Maternal-to-Zygotic Transition

In animal embryos, the maternal-to-zygotic transition (MZT) hands developmental control from maternal to zygotic gene products. We show that the maternal proteome represents more than half of the protein-coding capacity of Drosophila melanogaster’s genome, and that 2% of this proteome is rapidly degraded during the MZT. Cleared proteins include the post-transcriptional repressors Cup, Trailer hitch (TRAL), Maternal expression at 31B (ME31B), and Smaug (SMG). Although the ubiquitin-proteasome system is necessary for clearance of these repressors, distinct E3 ligase complexes target them: the C-terminal to Lis1 Homology (CTLH) complex targets Cup, TRAL, and ME31B for degradation early in the MZT and the Skp/Cullin/F-box-containing (SCF) complex targets SMG at the end of the MZT. Deleting the C-terminal 233 amino acids of SMG abrogates F-box protein interaction and confers immunity to degradation. Persistent SMG downregulates zygotic re-expression of mRNAs whose maternal contribution is degraded by SMG. Thus, clearance of SMG permits an orderly MZT.

Cao et al. show that 2% of the proteome is degraded in early Drosophila embryos, including a repressive ribonucleoprotein complex. Two E3 ubiquitin ligases separately act on distinct components of this complex to phase their clearance. Failure to degrade a key component, the Smaug RNA-binding protein, disrupts an orderly maternal-to-zygotic transition.

Triqler for MaxQuant: Enhancing Results from MaxQuant by Bayesian Error Propagation and Integration

Error estimation for differential protein quantification by label-free shotgun proteomics is challenging due to the multitude of error sources, each contributing uncertainty to the final results. We have previously designed a Bayesian model, Triqler, to combine such error terms into one combined quantification error. Here we present an interface for Triqler that takes MaxQuant results as input, allowing quick reanalysis of already processed data. We demonstrate that Triqler outperforms the original processing for a large set of both engineered and clinical/biological relevant data sets. Triqler and its interface to MaxQuant are available as a Python module under an Apache 2.0 license from https://pypi.org/project/triqler/.

Defining the carrier proteome limit for single-cell proteomics

Single-cell proteomics by mass spectrometry (SCoPE-MS) is a recently introduced method to quantify multiplexed single-cell proteomes. While this technique has generated great excitement, the underlying technologies (isobaric labeling and mass spectrometry (MS)) have technical limitations with the potential to affect data quality and biological interpretation. These limitations are particularly relevant when a carrier proteome, a sample added at 25-500× the amount of a single-cell proteome, is used to enable peptide identifications. Here we perform controlled experiments with increasing carrier proteome amounts and evaluate quantitative accuracy, as it relates to mass analyzer dynamic range, multiplexing level and number of ions sampled. We demonstrate that an increase in carrier proteome level requires a concomitant increase in the number of ions sampled to maintain quantitative accuracy. Lastly, we introduce Single-Cell Proteomics Companion (SCPCompanion), a software tool that enables rapid evaluation of single-cell proteomic data and recommends instrument and data analysis parameters for improved data quality.

The gain-of-function allele bamA E470K bypasses the essential requirement for BamD in β-barrel outer membrane protein assembly

The outer membrane (OM) of gram-negative bacteria confers innate resistance to toxins and antibiotics. Integral β-barrel outer membrane proteins (OMPs) function to establish and maintain the selective permeability of the OM. OMPs are assembled into the OM by the β-barrel assembly machine (BAM), which is composed of one OMP-BamA-and four lipoproteins-BamB, C, D, and E. BamB, C, and E can be removed individually with only minor effects on barrier function; however, depletion of either BamA or BamD causes a global defect in OMP assembly and results in cell death. We have identified a gain-of-function mutation, bamA _E470K , that bypasses the requirement for BamD. Although bamD::kan bamA _E470K cells exhibit growth and OM barrier defects, they assemble OMPs with surprising robustness. Our results demonstrate that BamD does not play a catalytic role in OMP assembly, but rather functions to regulate the activity of BamA.

A Bayesian Null Interval Hypothesis Test Controls False Discovery Rates and Improves Sensitivity in Label-Free Quantitative Proteomics

Statistical significance tests are a common feature in quantitative proteomics workflows. The Student's t-test is widely used to compute the statistical significance of a protein's change between two groups of samples. However, the t-test's null hypothesis asserts that the difference in means between two groups is exactly zero, often marking small but uninteresting fold-changes as statistically significant. Compensations to address this issue are widely used in quantitative proteomics, but we suggest that a replacement of the t-test with a Bayesian approach offers a better path forward. In this article, we describe a Bayesian hypothesis test in which the null hypothesis is an interval rather than a single point at zero; the width of the interval is estimated from population statistics. The improved sensitivity of the method substantially increases the number of truly changing proteins detected in two benchmark data sets (ProteomeXchange identifiers PXD005590 and PXD016470). The method has been implemented within FlashLFQ, an open-source software program that quantifies bottom-up proteomics search results obtained from any search tool. FlashLFQ is rapid, sensitive, and accurate and is available both as an easy-to-use graphical user interface (Windows) and as a command-line tool (Windows/Linux/OSX).

Multi-Omic Analyses Provide Links between Low-Dose Antibiotic Treatment and Induction of Secondary Metabolism in Burkholderia thailandensis

Low doses of antibiotics can trigger secondary metabolite biosynthesis in bacteria, but the underlying mechanisms are generally unknown. We sought to better understand this phenomenon by studying how the antibiotic trimethoprim activates the synthesis of the virulence factor malleilactone in Burkholderia thailandensis Using transcriptomics, quantitative multiplexed proteomics, and primary metabolomics, we systematically mapped the changes induced by trimethoprim. Surprisingly, even subinhibitory doses of the antibiotic resulted in broad transcriptional and translational alterations, with ∼8.5% of the transcriptome and ∼5% of the proteome up- or downregulated >4-fold. Follow-up studies with genetic-biochemical experiments showed that the induction of malleilactone synthesis can be sufficiently explained by the accumulation of methionine biosynthetic precursors, notably homoserine, as a result of inhibition of the folate pathway. Homoserine activated the malleilactone gene cluster via the transcriptional regulator MalR and gave rise to a secondary metabolome which was very similar to that generated by trimethoprim. Our work highlights the expansive changes that low-dose trimethoprim induces on bacterial physiology and provides insights into its stimulatory effect on secondary metabolism.IMPORTANCE The discovery of antibiotics ranks among the most significant accomplishments of the last century. Although the targets of nearly all clinical antibiotics are known, our understanding regarding their natural functions and the effects of subinhibitory concentrations is in its infancy. Stimulatory rather than inhibitory functions have been attributed to low-dose antibiotics. Among these, we previously found that antibiotics activate silent biosynthetic genes and thereby enhance the metabolic output of bacteria. The regulatory circuits underlying this phenomenon are unknown. We take a first step toward elucidating these circuits and show that low doses of trimethoprim (Tmp) have cell-wide effects on the saprophyte Burkholderia thailandensis Most importantly, inhibition of one-carbon metabolic processes by Tmp leads to an accumulation of homoserine, which induces the production of an otherwise silent cytotoxin via a LuxR-type transcriptional regulator. These results provide a starting point for uncovering the molecular basis of the hormetic effects of antibiotics.

YAP regulates cell size and growth dynamics via non-cell autonomous mediators

The Hippo pathway regulates organ size, regeneration, and cell growth by controlling the stability of the transcription factor, YAP (Yorkie in Drosophila). When there is tissue damage, YAP is activated allowing the restoration of homeostatic tissue size. The exact signals by which YAP is activated are still not fully understood, but its activation is known to affect both cell size and cell number. Here we used cultured cells to examine the coordinated regulation of cell size and number under the control of YAP. Our experiments in isogenic HEK293 cells reveal that YAP can affect cell size and number by independent circuits. Some of these effects are cell autonomous, such as proliferation, while others are mediated by secreted signals. In particular CYR61, a known secreted YAP target, is a non-cell autonomous mediator of cell survival, while another unidentified secreted factor controls cell size.

The Synthetic Phenotype of ΔbamB ΔbamE Double Mutants Results from a Lethal Jamming of the Bam Complex by the Lipoprotein RcsF

The selective permeability of the Gram-negative outer membrane (OM) is maintained by integral β-barrel outer membrane proteins (OMPs). The heteropentomeric β-barrel assembly machine (Bam) folds and inserts OMPs into the OM. Coordination of the essential proteins BamA and BamD is critical for OMP assembly and therefore the viability of the cell. The role of the nonessential lipoproteins BamBCE has yet to be characterized; however, genetic evidence suggests that they have nonoverlapping roles in OMP assembly. In this work, we quantify changes of the proteome in the conditional lethal ΔbamB ΔbamE double mutant. We show that cells lacking BamB and BamE have a global OMP defect that is a result of a lethal obstruction of an assembly-competent Bam complex by the lipoprotein RcsF. RcsF is a stress-sensing lipoprotein that is threaded through the lumen of abundant β-barrel OMPs by the Bam complex to expose the amino terminus on the cell surface. We demonstrate that simply removing this lipoprotein corrects the severe OMP assembly defect of the double mutant nearly as efficiently as a previously isolated suppressor mutation in bamA We propose that BamB and BamE play crucial, nonoverlapping roles to coordinate the activities of BamA and BamD during OMP biogenesis.IMPORTANCE Protein assembly into lipid bilayers is an essential process that ensures the viability of diverse organisms. In Gram-negative bacteria, the heteropentomeric β-barrel assembly machine (Bam) folds and inserts proteins into the outer membrane. Due to its essentiality, outer membrane protein (OMP) assembly by the Bam complex is an attractive target for antibiotic development. Here, we show that the conditional lethal phenotype of a mutant lacking two of the three nonessential lipoproteins, BamB and BamE, is caused by lethal jamming of the stripped-down Bam complex by a normally surface-exposed lipoprotein, RcsF. The heterotrimeric Bam complex (BamA, BamD, BamC) is nearly as efficient as the wild-type complex in OMP assembly if RcsF is removed. Our study highlights the importance of BamB and BamE in regulating the interaction between BamA and BamD and expands our understanding of the role of the Bam complex in outer membrane biogenesis.