A Scalable Approach for Protein False Discovery Rate Estimation in Large Proteomic Data Sets

Calculating the number of confidently identified proteins and estimating false discovery rate (FDR) is a challenge when analyzing very large proteomic data sets such as entire human proteomes. Biological and technical heterogeneity in proteomic experiments further add to the challenge and there are strong differences in opinion regarding the conceptual validity of a protein FDR and no consensus regarding the methodology for protein FDR determination. There are also limitations inherent to the widely used classic target–decoy strategy that particularly show when analyzing very large data sets and that lead to a strong over-representation of decoy identifications. In this study, we investigated the merits of the classic, as well as a novel target–decoy-based protein FDR estimation approach, taking advantage of a heterogeneous data collection comprised of ∼19,000 LC-MS/MS runs deposited in ProteomicsDB (https://www.proteomicsdb.org). The “picked” protein FDR approach treats target and decoy sequences of the same protein as a pair rather than as individual entities and chooses either the target or the decoy sequence depending on which receives the highest score. We investigated the performance of this approach in combination with q-value based peptide scoring to normalize sample-, instrument-, and search engine-specific differences. The “picked” target–decoy strategy performed best when protein scoring was based on the best peptide q-value for each protein yielding a stable number of true positive protein identifications over a wide range of q-value thresholds. We show that this simple and unbiased strategy eliminates a conceptual issue in the commonly used “classic” protein FDR approach that causes overprediction of false-positive protein identification in large data sets. The approach scales from small to very large data sets without losing performance, consistently increases the number of true-positive protein identifications and is readily implemented in proteomics analysis software.Shotgun proteomics is the most popular approach for large-scale identification and quantification of proteins. The rapid evolution of high-end mass spectrometers in recent years (1–5) has made proteomic studies feasible that identify and quantify as many as 10,000 proteins in a sample (6–8) and enables many lines of new scientific research including, for example, the analysis of many human proteomes, and proteome-wide protein–drug interaction studies (9–11). One fundamental step in most proteomic experiments is the identification of proteins in the biological system under investigation. To achieve this, proteins are digested into peptides, analyzed by LC-MS/MS, and tandem mass spectra are used to interrogate protein sequence databases using search engines that match experimental data to data generated in silico (12, 13). Peptide spectrum matches (PSMs)¹ are commonly assigned by a search engine using either a heuristic or a probabilistic scoring scheme (14–18). Proteins are then inferred from identified peptides and a protein score or a probability derived as a measure for the confidence in the identification (13, 19).Estimating the proportion of false matches (false discovery rate; FDR) in an experiment is important to assess and maintain the quality of protein identifications. Owing to its conceptual and practical simplicity, the most widely used strategy to estimate FDR in proteomics is the target–decoy database search strategy (target–decoy strategy; TDS) (20). The main assumption underlying this idea is that random matches (false positives) should occur with similar likelihood in the target database and the decoy (reversed, shuffled, or otherwise randomized) version of the same database (21, 22). The number of matches to the decoy database, therefore, provides an estimate of the number of random matches one should expect to obtain in the target database. The number of target and decoy hits can then be used to calculate either a local or a global FDR for a given data set (21–26). This general idea can be applied to control the FDR at the level of PSMs, peptides, and proteins, typically by counting the number of target and decoy observations above a specified score.Despite the significant practical impact of the TDS, it has been observed that a peptide FDR that results in an acceptable protein FDR (of say 1%) for a small or medium sized data set, turns into an unacceptably high protein FDR when the data set grows larger (22, 27). This is because the basic assumption of the classical TDS is compromised when a large proportion of the true positive proteins have already been identified. In small data sets, containing say only a few hundred to a few thousand proteins, random peptide matches will be distributed roughly equally over all decoy and “leftover” target proteins, allowing for a reasonably accurate estimation of false positive target identifications by using the number of decoy identifications. However, in large experiments comprising hundreds to thousands of LC-MS/MS runs, 10,000 or more target proteins may be genuinely and repeatedly identified, leaving an ever smaller number of (target) proteins to be hit by new false positive peptide matches. In contrast, decoy proteins are only hit by the occasional random peptide match but fully count toward the number of false positive protein identifications estimated from the decoy hits. The higher the number of genuinely identified target proteins gets, the larger this imbalance becomes. If this is not corrected for in the decoy space, an overestimation of false positives will occur.This problem has been recognized and e.g. Reiter and colleagues suggested a way for correcting for the overestimation of false positive protein hits termed MAYU (27). Following the main assumption that protein identifications containing false positive PSMs are uniformly distributed over the target database, MAYU models the number of false positive protein identifications using a hypergeometric distribution. Its parameters are estimated from the number of protein database entries and the total number of target and decoy protein identifications. The protein FDR is then estimated by dividing the number of expected false positive identifications (expectation value of the hypergeometric distribution) by the total number of target identifications. Although this approach was specifically designed for large data sets (tested on ∼1300 LC-MS/MS runs from digests of C. elegans proteins), it is not clear how far the approach actually scales. Another correction strategy for overestimation of false positive rates, the R factor, was suggested initially for peptides (28) and more recently for proteins (29). A ratio, R, of forward and decoy hits in the low probability range is calculated, where the number of true peptide or protein identifications is expected to be close to zero, and hence, R should approximate one. The number of decoy hits is then multiplied (corrected) by the R factor when performing FDR calculations. The approach is conceptually simpler than the MAYU strategy and easy to implement, but is also based on the assumption that the inflation of the decoy hits intrinsic in the classic target–decoy strategy occurs to the same extent in all probability ranges.In the context of the above, it is interesting to note that there is currently no consensus in the community regarding if and how protein FDRs should be calculated for data of any size. One perhaps extreme view is that, owing to issues and assumptions related to the peptide to protein inference step and ways of constructing decoy protein sequences, protein level FDRs cannot be meaningfully estimated at all (30). This is somewhat unsatisfactory as an estimate of protein level error in proteomic experiments is highly desirable. Others have argued that target–decoy searches are not even needed when accurate p values of individual PSMs are available (31) whereas others choose to tighten the PSM or peptide FDRs obtained from TDS analysis to whatever threshold necessary to obtain a desired protein FDR (32). This is likely too conservative.We have recently proposed an alternative protein FDR approach termed “picked” target–decoy strategy (picked TDS) that indicated improved performance over the classical TDS in a very large proteomic data set (9) but a systematic investigation of the idea had not been performed at the time. In this study, we further characterized the picked TDS for protein FDR estimation and investigated its scalability compared with that of the classic TDS FDR method in data sets of increasing size up to ∼19,000 LC-MS/MS runs. The results show that the picked TDS is effective in preventing decoy protein over-representation, identifies more true positive hits, and works equally well for small and large proteomic data sets.     

JUMP: A Tag-based Database Search Tool for Peptide Identification with High Sensitivity and Accuracy

Database search programs are essential tools for identifying peptides via mass spectrometry (MS) in shotgun proteomics. Simultaneously achieving high sensitivity and high specificity during a database search is crucial for improving proteome coverage. Here we present JUMP, a new hybrid database search program that generates amino acid tags and ranks peptide spectrum matches (PSMs) by an integrated score from the tags and pattern matching. In a typical run of liquid chromatography coupled with high-resolution tandem MS, more than 95% of MS/MS spectra can generate at least one tag, whereas the remaining spectra are usually too poor to derive genuine PSMs. To enhance search sensitivity, the JUMP program enables the use of tags as short as one amino acid. Using a target-decoy strategy, we compared JUMP with other programs (e.g. SEQUEST, Mascot, PEAKS DB, and InsPecT) in the analysis of multiple datasets and found that JUMP outperformed these preexisting programs. JUMP also permitted the analysis of multiple co-fragmented peptides from “mixture spectra” to further increase PSMs. In addition, JUMP-derived tags allowed partial de novo sequencing and facilitated the unambiguous assignment of modified residues. In summary, JUMP is an effective database search algorithm complementary to current search programs.Peptide identification by tandem mass spectra is a critical step in mass spectrometry (MS)-based¹ proteomics (1). Numerous computational algorithms and software tools have been developed for this purpose (2–6). These algorithms can be classified into three categories: (i) pattern-based database search, (ii) de novo sequencing, and (iii) hybrid search that combines database search and de novo sequencing. With the continuous development of high-performance liquid chromatography and high-resolution mass spectrometers, it is now possible to analyze almost all protein components in mammalian cells (7). In contrast to rapid data collection, it remains a challenge to extract accurate information from the raw data to identify peptides with low false positive rates (specificity) and minimal false negatives (sensitivity) (8).Database search methods usually assign peptide sequences by comparing MS/MS spectra to theoretical peptide spectra predicted from a protein database, as exemplified in SEQUEST (9), Mascot (10), OMSSA (11), X!Tandem (12), Spectrum Mill (13), ProteinProspector (14), MyriMatch (15), Crux (16), MS-GFDB (17), Andromeda (18), BaMS² (19), and Morpheus (20). Some other programs, such as SpectraST (21) and Pepitome (22), utilize a spectral library composed of experimentally identified and validated MS/MS spectra. These methods use a variety of scoring algorithms to rank potential peptide spectrum matches (PSMs) and select the top hit as a putative PSM. However, not all PSMs are correctly assigned. For example, false peptides may be assigned to MS/MS spectra with numerous noisy peaks and poor fragmentation patterns. If the samples contain unknown protein modifications, mutations, and contaminants, the related MS/MS spectra also result in false positives, as their corresponding peptides are not in the database. Other false positives may be generated simply by random matches. Therefore, it is of importance to remove these false PSMs to improve dataset quality. One common approach is to filter putative PSMs to achieve a final list with a predefined false discovery rate (FDR) via a target-decoy strategy, in which decoy proteins are merged with target proteins in the same database for estimating false PSMs (23–26). However, the true and false PSMs are not always distinguishable based on matching scores. It is a problem to set up an appropriate score threshold to achieve maximal sensitivity and high specificity (13, 27, 28).De novo methods, including Lutefisk (29), PEAKS (30), NovoHMM (31), PepNovo (32), pNovo (33), Vonovo (34), and UniNovo (35), identify peptide sequences directly from MS/MS spectra. These methods can be used to derive novel peptides and post-translational modifications without a database, which is useful, especially when the related genome is not sequenced. High-resolution MS/MS spectra greatly facilitate the generation of peptide sequences in these de novo methods. However, because MS/MS fragmentation cannot always produce all predicted product ions, only a portion of collected MS/MS spectra have sufficient quality to extract partial or full peptide sequences, leading to lower sensitivity than achieved with the database search methods.To improve the sensitivity of the de novo methods, a hybrid approach has been proposed to integrate peptide sequence tags into PSM scoring during database searches (36). Numerous software packages have been developed, such as GutenTag (37), InsPecT (38), Byonic (39), DirecTag (40), and PEAKS DB (41). These methods use peptide tag sequences to filter a protein database, followed by error-tolerant database searching. One restriction in most of these algorithms is the requirement of a minimum tag length of three amino acids for matching protein sequences in the database. This restriction reduces the sensitivity of the database search, because it filters out some high-quality spectra in which consecutive tags cannot be generated.In this paper, we describe JUMP, a novel tag-based hybrid algorithm for peptide identification. The program is optimized to balance sensitivity and specificity during tag derivation and MS/MS pattern matching. JUMP can use all potential sequence tags, including tags consisting of only one amino acid. When we compared its performance to that of two widely used search algorithms, SEQUEST and Mascot, JUMP identified ∼30% more PSMs at the same FDR threshold. In addition, the program provides two additional features: (i) using tag sequences to improve modification site assignment, and (ii) analyzing co-fragmented peptides from mixture MS/MS spectra.     

A Non-parametric Cutout Index for Robust Evaluation of Identified Proteins

This paper proposes a novel, automated method for evaluating sets of proteins identified using mass spectrometry. The remaining peptide-spectrum match score distributions of protein sets are compared to an empirical absent peptide-spectrum match score distribution, and a Bayesian non-parametric method reminiscent of the Dirichlet process is presented to accurately perform this comparison. Thus, for a given protein set, the process computes the likelihood that the proteins identified are correctly identified. First, the method is used to evaluate protein sets chosen using different protein-level false discovery rate (FDR) thresholds, assigning each protein set a likelihood. The protein set assigned the highest likelihood is used to choose a non-arbitrary protein-level FDR threshold. Because the method can be used to evaluate any protein identification strategy (and is not limited to mere comparisons of different FDR thresholds), we subsequently use the method to compare and evaluate multiple simple methods for merging peptide evidence over replicate experiments. The general statistical approach can be applied to other types of data (e.g. RNA sequencing) and generalizes to multivariate problems.Mass spectrometry is the predominant tool for characterizing complex protein mixtures. Using mass spectrometry, a heterogeneous protein sample is digested into peptides, which are separated by various features (e.g. retention time and mass-to-charge ratio), and fragmented to produce a large collection of spectra; these fragmentation spectra are matched to peptide sequences, and the peptide-spectrum matches (PSMs)¹ are scored (1). PSM scores from different peptide search engines and replicate experiments can be assembled to produce consensus scores for each peptide (2, 3). These peptide search results are then used to identify proteins (4).Inferring the protein content from these fragment ion spectra is difficult, and statistical methods have been developed with that goal. Protein identification methods (5–8) rank proteins according to the probability of their being present in the sample. Complementary target-decoy methods evaluate the proteins identified by searching fragmentation spectra against proteins that might be present (targets) and proteins that are absent (decoys). An identified target protein counts as a correct identification (increasing the estimated sensitivity), whereas each identified decoy protein counts as an incorrect identification (lowering the estimated specificity).Current target-decoy methods estimate the protein-level false discovery rate (FDR) for a set of identified proteins (9, 10), as well as the sensitivity at a particular arbitrary FDR threshold (11); however, these methods have two main shortcomings.First, current methods introduce strong statistical biases, which can be conservative (10) or optimistic (12) in different settings. These biases make current approaches unreliable for comparing different identification methods, because they implicitly favor methods that use similar assumptions. Automated evaluation tools that can be run without user-defined parameters are necessary in order to compare and improve existing analysis tools (13).Second, existing evaluation methods do not produce a single quality measure; instead, they estimate both FDR and sensitivity (which is estimated using the “absolute sensitivity,” which treats all targets as present and counts them as true identifications). For data sets with known protein contents (e.g. the protein standard data set considered), the absolute sensitivity is estimable; however, for more complex data sets with unknown contents, the measurement indicates the relative sensitivity. Even if one ignores statistical biases, there is currently no method for choosing a non-arbitrary FDR threshold, and it is currently not possible to decide which protein set is superior—one with a lower sensitivity and stricter FDR, or another with a higher sensitivity and less stringent FDR. The former is currently favored but might result in significant information loss. Arbitrary thresholds have significant effects: in the yeast data analyzed, 1% and 5% FDR thresholds, respectively, yielded 1289 and 1570 identified protein groups (grouping is discussed in the supplementary “Methods” section). Even with such a simple data set, this subtle change results in 281 more target identifications, of which unknown subsets of 66 (0.05 × 1570 − 0.01 × 1289 ≈ 66) are expected to be false identifications and 215 are expected to be true identifications (281 − 66 = 215).Here we introduce the non-parametric cutout index (npCI), a novel, automated target-decoy method that can be used to compute a single robust and parameter-free quality measure for protein identifications. Our method does not require prior expertise in order for the user to select parameters or run the computation. The npCI employs target-decoy analysis at the PSM level, where its assumptions are more applicable (4). Rather than use assumptions to model PSM scores matching present proteins, our method remains agnostic to the characteristics of present proteins and analyzes PSMs not explained by the identified proteins. If the correct present set of proteins is known, then the distribution of remaining, unexplained PSM scores resembles the decoy distribution (14). We extend this idea and present a general graphical framework to evaluate a set of protein identifications by computing the likelihood that the remaining PSMs and decoy PSMs are drawn from the same distribution (Fig. 1).Open in a separate windowFig. 1.Schematic for non-parametric probabilistic evaluation of identified proteins. Under the supposition that the identified protein set (blue) is present, all peptides matching those proteins (also blue) might be present and have an unknown score distribution. When the correct set of proteins is identified, the remaining peptides (i.e. those not matching any shaded proteins in this figure) have a score distribution resembling that of absent peptides. Thus, the similarity of the remaining peptide score distribution (red dashed line) to the absent peptide score distribution (black solid line) determines the quality of the identified proteins.Existing non-parametric statistical tests evaluating the similarity between two collections of samples (e.g. Kolmogorov–Smirnov test, used in Ref. 14, and the Wilcoxon signed rank test) were inadequate because infrequent but significant outliers (e.g. high-scoring PSMs) are largely ignored by these methods. Likewise, information-theoretic measures, such as the Kullback–Leibler divergence, were inadequate because they require a prior on the smoothing parameter that weighs more smoothing and higher similarity against less smoothing and lower similarity (a problem reminiscent of the original compromise between sensitivity and FDR); without the application of such a prior, the optimal Kullback–Leibler divergence occurs with infinite smoothing, which will make any distributions equal, rendering them completely uninformative and thus making it impossible to distinguish one identified protein set from another. For these reasons, we derived a novel, Bayesian, non-parametric process to compute the likelihood that two continuous collections are drawn from the same distribution. It can be used to provide a robust and efficient evaluation of discoveries.     

Maximizing Peptide Identification Events in Proteomic Workflows Using Data-Dependent Acquisition (DDA)

Current analytical strategies for collecting proteomic data using data-dependent acquisition (DDA) are limited by the low analytical reproducibility of the method. Proteomic discovery efforts that exploit the benefits of DDA, such as providing peptide sequence information, but that enable improved analytical reproducibility, represent an ideal scenario for maximizing measureable peptide identifications in “shotgun”-type proteomic studies. Therefore, we propose an analytical workflow combining DDA with retention time aligned extracted ion chromatogram (XIC) areas obtained from high mass accuracy MS1 data acquired in parallel. We applied this workflow to the analyses of sample matrixes prepared from mouse blood plasma and brain tissues and observed increases in peptide detection of up to 30.5% due to the comparison of peptide MS1 XIC areas following retention time alignment of co-identified peptides. Furthermore, we show that the approach is quantitative using peptide standards diluted into a complex matrix. These data revealed that peptide MS1 XIC areas provide linear response of over three orders of magnitude down to low femtomole (fmol) levels. These findings argue that augmenting “shotgun” proteomic workflows with retention time alignment of peptide identifications and comparative analyses of corresponding peptide MS1 XIC areas improve the analytical performance of global proteomic discovery methods using DDA.Label-free methods in mass spectrometry-based proteomics, such as those used in common “shotgun” proteomic studies, provide peptide sequence information as well as relative measurements of peptide abundance (1–3). A common data acquisition strategy is based on data-dependent acquisition (DDA)¹ where the most abundant precursor ions are selected for tandem mass spectrometry (MS/MS) analysis (1–2). DDA attempts to minimize redundant peptide precursor selection and maximize the depth of proteome coverage (2). However, the analytical reproducibility of peptide identifications obtained using DDA-based methods result in <75% overlap between technical replicates (3–4). Comparisons of peptide identifications between replicate analyses have shown that the rate of new peptide identifications increases sharply following two replicate sample injections and gradually tapers off after approximately five replicate injections (4). This phenomenon is due, in part, to the semirandom sampling of peptides in a DDA experiment (5).Alternate label-free methods focused on measuring the abundance of intact peptide ions, such as the accurate mass and time tag (AMT) approach (6–8, 42), are aimed at differential analyses of extracted ion chromatogram (XIC) areas integrated from high mass accuracy peptide precursor mass spectra (MS1 spectra) exhibiting discrete chromatographic elution times. This method is particularly powerful for the analysis of post-translationally modified (PTM) peptides as pairing the low abundance of PTM candidates with the variable nature of DDA complicates comparisons between samples (9, 43). However, label-free strategies focused on the analysis of peptide MS1 XIC areas are dependent on a priori knowledge of peptide ions and retention times (2–10). Thus, prospective analyses of samples are needed to assess peptides and their respective retention times. This prospective analysis may not be possible for reagent-limited samples. Further, the usage of previously established peptide features in the analysis of different sample types can be confounded by unknown matrix effects that can produce variable retention time characteristics and peptide ion suppression (2). Therefore, proteomic strategies that make use of DDA, to provide peptide sequence information and identify features within the sample, but that also use MS1 data for comparisons between samples, represent an ideal combination for maximizing measureable peptide identification events in “shotgun” proteomic discovery analyses.Here we describe an analytical workflow that combines traditional DDA methods with the analysis of retention time aligned XIC areas extracted from high mass accuracy peptide precursor MS1 spectra. This method resulted in a 25.1% (±6.6%) increase in measureable peptide identification events across samples of diverse composition because of the inferential extraction of peptide MS1 XIC areas in sample sets lacking corresponding MS/MS events. These findings were observed in measurements of peptide MS1 XIC abundances using sample types ranging from tryptic digests of olfactory bulb tissues dissected from Homer2 knockout and wild-type mice to mouse blood plasma exhibiting differential levels of hemolysis. We further establish that this method is quantitative using a dilution series of known bovine standard peptide concentrations spiked into mouse blood plasma. These data show that comparative analysis between samples should be performed using peptide MS1 data as opposed to semirandomly sampled peptide MS/MS data. This approach maximizes the number of peptides that can be compared between samples.     

Analysis of Free Energy Signals Arising from Nucleotide Hybridization Between rRNA and mRNA Sequences during Translation in Eubacteria

A decoding algorithm is tested that mechanistically models the progressive alignments that arise as the mRNA moves past the rRNA tail during translation elongation. Each of these alignments provides an opportunity for hybridization between the single-stranded, -terminal nucleotides of the 16S rRNA and the spatially accessible window of mRNA sequence, from which a free energy value can be calculated. Using this algorithm we show that a periodic, energetic pattern of frequency 1/3 is revealed. This periodic signal exists in the majority of coding regions of eubacterial genes, but not in the non-coding regions encoding the 16S and 23S rRNAs. Signal analysis reveals that the population of coding regions of each bacterial species has a mean phase that is correlated in a statistically significant way with species () content. These results suggest that the periodic signal could function as a synchronization signal for the maintenance of reading frame and that codon usage provides a mechanism for manipulation of signal phase.[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32]     

Algorithms for Finding Small Attractors in Boolean Networks

A Boolean network is a model used to study the interactions between different genes in genetic regulatory networks. In this paper, we present several algorithms using gene ordering and feedback vertex sets to identify singleton attractors and small attractors in Boolean networks. We analyze the average case time complexities of some of the proposed algorithms. For instance, it is shown that the outdegree-based ordering algorithm for finding singleton attractors works in time for , which is much faster than the naive time algorithm, where is the number of genes and is the maximum indegree. We performed extensive computational experiments on these algorithms, which resulted in good agreement with theoretical results. In contrast, we give a simple and complete proof for showing that finding an attractor with the shortest period is NP-hard.[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32]     

Conserved Peptide Fragmentation as a Benchmarking Tool for Mass Spectrometers and a Discriminating Feature for Targeted Proteomics

Quantifying the similarity of spectra is an important task in various areas of spectroscopy, for example, to identify a compound by comparing sample spectra to those of reference standards. In mass spectrometry based discovery proteomics, spectral comparisons are used to infer the amino acid sequence of peptides. In targeted proteomics by selected reaction monitoring (SRM) or SWATH MS, predetermined sets of fragment ion signals integrated over chromatographic time are used to identify target peptides in complex samples. In both cases, confidence in peptide identification is directly related to the quality of spectral matches. In this study, we used sets of simulated spectra of well-controlled dissimilarity to benchmark different spectral comparison measures and to develop a robust scoring scheme that quantifies the similarity of fragment ion spectra. We applied the normalized spectral contrast angle score to quantify the similarity of spectra to objectively assess fragment ion variability of tandem mass spectrometric datasets, to evaluate portability of peptide fragment ion spectra for targeted mass spectrometry across different types of mass spectrometers and to discriminate target assays from decoys in targeted proteomics. Altogether, this study validates the use of the normalized spectral contrast angle as a sensitive spectral similarity measure for targeted proteomics, and more generally provides a methodology to assess the performance of spectral comparisons and to support the rational selection of the most appropriate similarity measure. The algorithms used in this study are made publicly available as an open source toolset with a graphical user interface.In “bottom-up” proteomics, peptide sequences are identified by the information contained in their fragment ion spectra (1). Various methods have been developed to generate peptide fragment ion spectra and to match them to their corresponding peptide sequences. They can be broadly grouped into discovery and targeted methods. In the widely used discovery (also referred to as shotgun) proteomic approach, peptides are identified by establishing peptide to spectrum matches via a method referred to as database searching. Each acquired fragment ion spectrum is searched against theoretical peptide fragment ion spectra computed from the entries of a specified sequence database, whereby the database search space is constrained to a user defined precursor mass tolerance (2, 3). The quality of the match between experimental and theoretical spectra is typically expressed with multiple scores. These include the number of matching or nonmatching fragments, the number of consecutive fragment ion matches among others. With few exceptions (4–7) commonly used search engines do not use the relative intensities of the acquired fragment ion signals even though this information could be expected to strengthen the confidence of peptide identification because the relative fragment ion intensity pattern acquired under controlled fragmentation conditions can be considered as a unique “fingerprint” for a given precursor. Thanks to community efforts in acquiring and sharing large number of datasets, the proteomes of some species are now essentially mapped out and experimental fragment ion spectra covering entire proteomes are increasingly becoming accessible through spectral databases (8–16). This has catalyzed the emergence of new proteomics strategies that differ from classical database searching in that they use prior spectral information to identify peptides. Those comprise inclusion list sequencing (directed sequencing), spectral library matching, and targeted proteomics (17). These methods explicitly use the information contained in empirical fragment ion spectra, including the fragment ion signal intensity to identify the target peptide. For these methods, it is therefore of highest importance to accurately control and quantify the degree of reproducibility of the fragment ion spectra across experiments, instruments, labs, methods, and to quantitatively assess the similarity of spectra. To date, dot product (18–24), its corresponding arccosine spectral contrast angle (25–27) and (Pearson-like) spectral correlation (28–31), and other geometrical distance measures (18, 32), have been used in the literature for assessing spectral similarity. These measures have been used in different contexts including shotgun spectra clustering (19, 26), spectral library searching (18, 20, 21, 24, 25, 27–29), cross-instrument fragmentation comparisons (22, 30) and for scoring transitions in targeted proteomics analyses such as selected reaction monitoring (SRM)¹ (23, 31). However, to our knowledge, those scores have never been objectively benchmarked for their performance in discriminating well-defined levels of dissimilarities between spectra. In particular, similarity scores obtained by different methods have not yet been compared for targeted proteomics applications, where the sensitive discrimination of highly similar spectra is critical for the confident identification of targeted peptides.In this study, we have developed a method to objectively assess the similarity of fragment ion spectra. We provide an open-source toolset that supports these analyses. Using a computationally generated benchmark spectral library with increasing levels of well-controlled spectral dissimilarity, we performed a comprehensive and unbiased comparison of the performance of the main scores used to assess spectral similarity in mass spectrometry.We then exemplify how this method, in conjunction with its corresponding benchmarked perturbation spectra set, can be applied to answer several relevant questions for MS-based proteomics. As a first application, we show that it can efficiently assess the absolute levels of peptide fragmentation variability inherent to any given mass spectrometer. By comparing the instrument''s intrinsic fragmentation conservation distribution to that of the benchmarked perturbation spectra set, nominal values of spectral similarity scores can indeed be translated into a more directly understandable percentage of variability inherent to the instrument fragmentation. As a second application, we show that the method can be used to derive an absolute measure to estimate the conservation of peptide fragmentation between instruments or across proteomics methods. This allowed us to quantitatively evaluate, for example, the transferability of fragment ion spectra acquired by data dependent analysis in a first instrument into a fragment/transition assay list used for targeted proteomics applications (e.g. SRM or targeted extraction of data independent acquisition SWATH MS (33)) on another instrument. Third, we used the method to probe the fragmentation patterns of peptides carrying a post-translation modification (e.g. phosphorylation) by comparing the spectra of modified peptide with those of their unmodified counterparts. Finally, we used the method to determine the overall level of fragmentation conservation that is required to support target-decoy discrimination and peptide identification in targeted proteomics approaches such as SRM and SWATH MS.     

Preserved Proteins from Extinct Bison latifrons Identified by Tandem Mass Spectrometry; Hydroxylysine Glycosides are a Common Feature of Ancient Collagen

Bone samples from several vertebrates were collected from the Ziegler Reservoir fossil site, in Snowmass Village, Colorado, and processed for proteomics analysis. The specimens come from Pleistocene megafauna Bison latifrons, dating back ∼120,000 years. Proteomics analysis using a simplified sample preparation procedure and tandem mass spectrometry (MS/MS) was applied to obtain protein identifications. Several bioinformatics resources were used to obtain peptide identifications based on sequence homology to extant species with annotated genomes. With the exception of soil sample controls, all samples resulted in confident peptide identifications that mapped to type I collagen. In addition, we analyzed a specimen from the extinct B. latifrons that yielded peptide identifications mapping to over 33 bovine proteins. Our analysis resulted in extensive fibrillar collagen sequence coverage, including the identification of posttranslational modifications. Hydroxylysine glucosylgalactosylation, a modification thought to be involved in collagen fiber formation and bone mineralization, was identified for the first time in an ancient protein dataset. Meta-analysis of data from other studies indicates that this modification may be common in well-preserved prehistoric samples. Additional peptide sequences from extracellular matrix (ECM) and non-ECM proteins have also been identified for the first time in ancient tissue samples. These data provide a framework for analyzing ancient protein signatures in well-preserved fossil specimens, while also contributing novel insights into the molecular basis of organic matter preservation. As such, this analysis has unearthed common posttranslational modifications of collagen that may assist in its preservation over time. The data are available via ProteomeXchange with identifier PXD001827.During the last decade, paleontology and taphonomy (the study of decaying organisms over time and the fossilization processes) have begun to overlap with the field of proteomics to shed new light on preserved organic matter in fossilized bones (1–4). These bones represent a time capsule of ancient biomolecules, owing to their natural resistance to post mortem decay arising from a unique combination of mechanical, structural, and chemical properties (4–7).Although bones can be cursorily described as a composite of collagen (protein) and hydroxyapatite (mineral), fossilized bones undergo three distinct diagenesis pathways: (i) chemical deterioration of the organic phase; (ii) chemical deterioration of the mineral phase; and (iii) (micro)biological attack of the composite (6). In addition, the rate of these degradation pathways are affected by temperature, as higher burial temperatures have been shown to accelerate these processes (6, 8). Though relatively unusual, the first of these three pathways results in a slower deterioration process, which is more generally mitigated under (6) specific environmental constraints, such as geochemical stability (stable temperature and acidity) that promote bone mineral preservation. Importantly, slower deterioration results in more preserved biological materials that are more amenable to downstream analytical assays. One example of this is the controversial case of bone and soft-tissue preservation from the Cretaceous/Tertiary boundary (9–22). In light of these and other studies of ancient biomolecules, paleontological models have proposed that organic biomolecules in ancient samples, such as collagen sequences from the 80 million-year-(my)-old Campanian hadrosaur, Brachylophosaurus canadensis (16) or 68-my-old Tyrannosaurus rex, might be protected by the microenvironment within bones. Such spaces are believed to form a protective shelter that is able to reduce the effects of diagenetic events. In addition to collagen, preserved biomolecules include blood proteins, cellular lipids, and DNA (4, 5). While the maximum estimated lifespan of DNA in bones is ∼20,000 years (ky) at 10 °C, bone proteins have an even longer lifespan, making them an exceptional target for analysis to gain relevant insights into fossilized samples (6). Indeed, the survival of collagen, which is considered to be the most abundant bone protein, is estimated to be in the range 340 ky at 20 °C. Similarly, osteocalcin, the second-most abundant bone protein, can persist for ≈45 ky at 20 °C, thus opening an unprecedented analytical window to study extremely old samples (2, 4, 23).Although ancient DNA amplification and sequencing can yield interesting clues and potential artifacts from contaminating agents (7, 24–28), the improved preservation of ancient proteins provides access to a reservoir of otherwise unavailable genetic information for phylogenetic inference (25, 29, 30). In particular, mass spectrometry (MS)-based screening of species-specific collagen peptides has recently been used as a low-cost, rapid alternative to DNA sequencing for taxonomic attribution of morphologically unidentifiable small bone fragments and teeth stemming from diverse archeological contexts (25, 31–33).For over five decades, researchers have presented biochemical evidence for the existence of preserved protein material from ancient bone samples (34–36). One of the first direct measurements was by amino acid analysis, which showed that the compositional profile of ancient samples was consistent with collagens in modern bone samples (37–39). Preservation of organic biomolecules, either from bone, dentin, antlers, or ivory, has been investigated by radiolabeled ¹⁴C fossil dating (40) to provide an avenue of delineating evolutionary divergence from extant species (3, 41, 42). It is also important to note that these parameters primarily depend on ancient bone collagen as the levels remain largely unchanged (a high percentage of collagen is retained, as gleaned by laboratory experiments on bone taphonomy (6)). Additionally, antibody-based immunostaining methods have given indirect evidence of intact peptide amide bonds (43–45) to aid some of the first evidence of protein other than collagen and osteocalcin in ancient mammoth (43) and human specimens (46).In the past, mass spectrometry has been used to obtain MS signals consistent with modern osteocalcin samples (2, 47), and eventually postsource decay peptide fragmentation was used to confirm the identification of osteocalcin in fossil hominids dating back ∼75 ky (48). More recently, modern “bottom-up” proteomic methods were applied to mastodon and T. rex samples (10), complementing immunohistochemistry evidence (13, 17). The results hinted at the potential of identifying peptides from proteolytic digest of well-preserved bone samples. This work also highlighted the importance of minimizing sources of protein contamination and adhering to data publication guidelines (20, 21). In the past few years, a very well-preserved juvenile mammoth referred to as Lyuba was discovered in the Siberian permafrost and analyzed using high-resolution tandem mass spectrometry (29). This study was followed with a report by Wadsworth and Buckley (30) describing the analysis of proteins from 19 bovine bone samples spanning 4 ky to 1.5 my. Both of these groups reported the identification of additional collagen and noncollagen proteins.Recently, a series of large extinct mammal bones were unearthed at a reservoir near Snowmass Village, Colorado, USA (49, 50). The finding was made during a construction project at the Ziegler Reservoir, a fossil site that was originally a lake formed at an elevation of ∼2,705 m during the Bull Lake glaciations ∼140 ky ago (49, 51). The original lake area was ∼5 hectares in size with a total catchment of ∼14 hectares and lacked a direct water flow inlet or outlet. This closed drainage basin established a relatively unique environment that resulted in the exceptional preservation of plant material, insects (52), and vertebrate bones (49). In particular, a cranial specimen from extinct Bison latifrons was unearthed from the Biostratigraphic Zone/Marine Oxygen Isotope Stage (MIS) 5d, which dates back to ∼120 ky (53, 54).Here, we describe the use of paleoproteomics, for the identification of protein remnants with a focus on a particularly unique B. latifrons cranial specimen found at the Ziegler site. We developed a simplified sample processing approach that allows for analysis of low milligram quantities of ancient samples for peptide identification. Our method avoids the extensive demineralization steps of traditional protocols and utilizes an acid labile detergent to allow for efficient extraction and digestion without the need for additional sample cleanup steps. This approach was applied to a specimen from B. latifrons that displayed visual and mechanical properties consistent with the meninges, a fibrous tissue that lines the cranial cavity. Bioinformatics analysis revealed the presence of a recurring glycosylation signature in well-preserved collagens. In particular, the presence of glycosylated hydroxylysine residues was identified as a unique feature of bone fossil collagen, as gleaned through meta-analyses of raw data from previous reports on woolly mammoth (Mammuthus primigenius) and bovine samples (29, 30). The results from these meta-analyses indicate a common, unique feature of collagen that coincides with, and possibly contributes to its preservation.     

Proinsulin and the Genetics of Diabetes Mellitus

Insulin plays a central role in the regulation of vertebrate metabolism. The hormone, the post-translational product of a single-chain precursor, is a globular protein containing two chains, A (21 residues) and B (30 residues). Recent advances in human genetics have identified dominant mutations in the insulin gene causing permanent neonatal-onset DM² (1–4). The mutations are predicted to block folding of the precursor in the ER of pancreatic β-cells. Although expression of the wild-type allele would in other circumstances be sufficient to maintain homeostasis, studies of a corresponding mouse model (5–7) suggest that the misfolded variant perturbs wild-type biosynthesis (8, 9). Impaired β-cell secretion is associated with ER stress, distorted organelle architecture, and cell death (10). These findings have renewed interest in insulin biosynthesis (11–13) and the structural basis of disulfide pairing (14–19). Protein evolution is constrained not only by structure and function but also by susceptibility to toxic misfolding.Insulin plays a central role in the regulation of vertebrate metabolism. The hormone, the post-translational product of a single-chain precursor, is a globular protein containing two chains, A (21 residues) and B (30 residues). Recent advances in human genetics have identified dominant mutations in the insulin gene causing permanent neonatal-onset DM² (1–4). The mutations are predicted to block folding of the precursor in the ER of pancreatic β-cells. Although expression of the wild-type allele would in other circumstances be sufficient to maintain homeostasis, studies of a corresponding mouse model (5–7) suggest that the misfolded variant perturbs wild-type biosynthesis (8, 9). Impaired β-cell secretion is associated with ER stress, distorted organelle architecture, and cell death (10). These findings have renewed interest in insulin biosynthesis (11–13) and the structural basis of disulfide pairing (14–19). Protein evolution is constrained not only by structure and function but also by susceptibility to toxic misfolding.     

The Wavelet-Based Cluster Analysis for Temporal Gene Expression Data

A variety of high-throughput methods have made it possible to generate detailed temporal expression data for a single gene or large numbers of genes. Common methods for analysis of these large data sets can be problematic. One challenge is the comparison of temporal expression data obtained from different growth conditions where the patterns of expression may be shifted in time. We propose the use of wavelet analysis to transform the data obtained under different growth conditions to permit comparison of expression patterns from experiments that have time shifts or delays. We demonstrate this approach using detailed temporal data for a single bacterial gene obtained under 72 different growth conditions. This general strategy can be applied in the analysis of data sets of thousands of genes under different conditions.[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29]     

MixGF: Spectral Probabilities for Mixture Spectra from more than One Peptide

In large-scale proteomic experiments, multiple peptide precursors are often cofragmented simultaneously in the same mixture tandem mass (MS/MS) spectrum. These spectra tend to elude current computational tools because of the ubiquitous assumption that each spectrum is generated from only one peptide. Therefore, tools that consider multiple peptide matches to each MS/MS spectrum can potentially improve the relatively low spectrum identification rate often observed in proteomics experiments. More importantly, data independent acquisition protocols promoting the cofragmentation of multiple precursors are emerging as alternative methods that can greatly improve the throughput of peptide identifications but their success also depends on the availability of algorithms to identify multiple peptides from each MS/MS spectrum. Here we address a fundamental question in the identification of mixture MS/MS spectra: determining the statistical significance of multiple peptides matched to a given MS/MS spectrum. We propose the MixGF generating function model to rigorously compute the statistical significance of peptide identifications for mixture spectra and show that this approach improves the sensitivity of current mixture spectra database search tools by a ≈30–390%. Analysis of multiple data sets with MixGF reveals that in complex biological samples the number of identified mixture spectra can be as high as 20% of all the identified spectra and the number of unique peptides identified only in mixture spectra can be up to 35.4% of those identified in single-peptide spectra.The advancement of technology and instrumentation has made tandem mass (MS/MS)¹ spectrometry the leading high-throughput method to analyze proteins (1, 2, 3). In typical experiments, tens of thousands to millions of MS/MS spectra are generated and enable researchers to probe various aspects of the proteome on a large scale. Part of this success hinges on the availability of computational methods that can analyze the large amount of data generated from these experiments. The classical question in computational proteomics asks: given an MS/MS spectrum, what is the peptide that generated the spectrum? However, it is increasingly being recognized that this assumption that each MS/MS spectrum comes from only one peptide is often not valid. Several recent analyses show that as many as 50% of the MS/MS spectra collected in typical proteomics experiments come from more than one peptide precursor (4, 5). The presence of multiple peptides in mixture spectra can decrease their identification rate to as low as one half of that for MS/MS spectra generated from only one peptide (6, 7, 8). In addition, there have been numerous developments in data independent acquisition (DIA) technologies where multiple peptide precursors are intentionally selected to cofragment in each MS/MS spectrum (9, 10, 11, 12, 13, 14, 15). These emerging technologies can address some of the enduring disadvantages of traditional data-dependent acquisition (DDA) methods (e.g. low reproducibility (16)) and potentially increase the throughput of peptide identification 5–10 fold (4, 17). However, despite the growing importance of mixture spectra in various contexts, there are still only a few computational tools that can analyze mixture spectra from more than one peptide (18, 19, 20, 21, 8, 22). Our recent analysis indicated that current database search methods for mixture spectra still have relatively low sensitivity compared with their single-peptide counterpart and the main bottleneck is their limited ability to separate true matches from false positive matches (8). Traditionally problem of peptide identification from MS/MS spectra involves two sub-problems: 1) define a Peptide-Spectrum-Match (PSM) scoring function that assigns each MS/MS spectrum to the peptide sequence that most likely generated the spectrum; and 2) given a set of top-scoring PSMs, select a subset that corresponds to statistical significance PSMs. Here we focus on the second problem, which is still an ongoing research question even for the case of single-peptide spectra (23, 24, 25, 26). Intuitively the second problem is difficult because one needs to consider spectra across the whole data set (instead of comparing different peptide candidates against one spectrum as in the first problem) and PSM scoring functions are often not well-calibrated across different spectra (i.e. a PSM score of 50 may be good for one spectrum but poor for a different spectrum). Ideally, a scoring function will give high scores to all true PSMs and low scores to false PSMs regardless of the peptide or spectrum being considered. However, in practice, some spectra may receive higher scores than others simply because they have more peaks or their precursor mass results in more peptide candidates being considered from the sequence database (27, 28). Therefore, a scoring function that accounts for spectrum or peptide-specific effects can make the scores more comparable and thus help assess the confidence of identifications across different spectra. The MS-GF solution to this problem is to compute the per-spectrum statistical significance of each top-scoring PSM, which can be defined as the probability that a random peptide (out of all possible peptide within parent mass tolerance) will match to the spectrum with a score at least as high as that of the top-scoring PSM. This measures how good the current best match is in relation to all possible peptides matching to the same spectrum, normalizing any spectrum effect from the scoring function. Intuitively, our proposed MixGF approach extends the MS-GF approach to now calculate the statistical significance of the top pair of peptides matched from the database to a given mixture spectrum M (i.e. the significance of the top peptide–peptide spectrum match (PPSM)). As such, MixGF determines the probability that a random pair of peptides (out of all possible peptides within parent mass tolerance) will match a given mixture spectrum with a score at least as high as that of the top-scoring PPSM.Despite the theoretical attractiveness of computing statistical significance, it is generally prohibitive for any database search methods to score all possible peptides against a spectrum. Therefore, earlier works in this direction focus on approximating this probability by assuming the score distribution of all PSMs follows certain analytical form such as the normal, Poisson or hypergeometric distributions (29, 30, 31). In practice, because score distributions are highly data-dependent and spectrum-specific, these model assumptions do not always hold. Other approaches tried to learn the score distribution empirically from the data (29, 27). However, one is most interested in the region of the score distribution where only a small fraction of false positives are allowed (typically at 1% FDR). This usually corresponds to the extreme tail of the distribution where p values are on the order of 10⁻⁹ or lower and thus there is typically lack of sufficient data points to accurately model the tail of the score distribution (32). More recently, Kim et al. (24) and Alves et al. (33), in parallel, proposed a generating function approach to compute the exact score distribution of random peptide matches for any spectra without explicitly matching all peptides to a spectrum. Because it is an exact computation, no assumption is made about the form of score distribution and the tail of the distribution can be computed very accurately. As a result, this approach substantially improved the ability to separate true matches from false positive ones and lead to a significant increase in sensitivity of peptide identification over state-of-the-art database search tools in single-peptide spectra (24).For mixture spectra, it is expected that the scores for the top-scoring match will be even less comparable across different spectra because now more than one peptide and different numbers of peptides can be matched to each spectrum at the same time. We extend the generating function approach (24) to rigorously compute the statistical significance of multiple-Peptide-Spectrum Matches (mPSMs) and demonstrate its utility toward addressing the peptide identification problem in mixture spectra. In particular, we show how to extend the generating approach for mixture from two peptides. We focus on this relatively simple case of mixture spectra because it accounts for a large fraction of mixture spectra presented in traditional DDA workflows (5). This allows us to test and develop algorithmic concepts using readily-available DDA data because data with more complex mixture spectra such as those from DIA workflows (11) is still not widely available in public repositories.     

Splitting the BLOSUM Score into Numbers of Biological Significance

Mathematical tools developed in the context of Shannon information theory were used to analyze the meaning of the BLOSUM score, which was split into three components termed as the BLOSUM spectrum (or BLOSpectrum). These relate respectively to the sequence convergence (the stochastic similarity of the two protein sequences), to the background frequency divergence (typicality of the amino acid probability distribution in each sequence), and to the target frequency divergence (compliance of the amino acid variations between the two sequences to the protein model implicit in the BLOCKS database). This treatment sharpens the protein sequence comparison, providing a rationale for the biological significance of the obtained score, and helps to identify weakly related sequences. Moreover, the BLOSpectrum can guide the choice of the most appropriate scoring matrix, tailoring it to the evolutionary divergence associated with the two sequences, or indicate if a compositionally adjusted matrix could perform better.[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29]     

High-definition De Novo Sequencing of Crustacean Hyperglycemic Hormone (CHH)-family Neuropeptides

A complete understanding of the biological functions of large signaling peptides (>4 kDa) requires comprehensive characterization of their amino acid sequences and post-translational modifications, which presents significant analytical challenges. In the past decade, there has been great success with mass spectrometry-based de novo sequencing of small neuropeptides. However, these approaches are less applicable to larger neuropeptides because of the inefficient fragmentation of peptides larger than 4 kDa and their lower endogenous abundance. The conventional proteomics approach focuses on large-scale determination of protein identities via database searching, lacking the ability for in-depth elucidation of individual amino acid residues. Here, we present a multifaceted MS approach for identification and characterization of large crustacean hyperglycemic hormone (CHH)-family neuropeptides, a class of peptide hormones that play central roles in the regulation of many important physiological processes of crustaceans. Six crustacean CHH-family neuropeptides (8–9.5 kDa), including two novel peptides with extensive disulfide linkages and PTMs, were fully sequenced without reference to genomic databases. High-definition de novo sequencing was achieved by a combination of bottom-up, off-line top-down, and on-line top-down tandem MS methods. Statistical evaluation indicated that these methods provided complementary information for sequence interpretation and increased the local identification confidence of each amino acid. Further investigations by MALDI imaging MS mapped the spatial distribution and colocalization patterns of various CHH-family neuropeptides in the neuroendocrine organs, revealing that two CHH-subfamilies are involved in distinct signaling pathways.Neuropeptides and hormones comprise a diverse class of signaling molecules involved in numerous essential physiological processes, including analgesia, reward, food intake, learning and memory (1). Disorders of the neurosecretory and neuroendocrine systems influence many pathological processes. For example, obesity results from failure of energy homeostasis in association with endocrine alterations (2, 3). Previous work from our lab used crustaceans as model organisms found that multiple neuropeptides were implicated in control of food intake, including RFamides, tachykinin related peptides, RYamides, and pyrokinins (4–6).Crustacean hyperglycemic hormone (CHH)¹ family neuropeptides play a central role in energy homeostasis of crustaceans (7–17). Hyperglycemic response of the CHHs was first reported after injection of crude eyestalk extract in crustaceans. Based on their preprohormone organization, the CHH family can be grouped into two sub-families: subfamily-I containing CHH, and subfamily-II containing molt-inhibiting hormone (MIH) and mandibular organ-inhibiting hormone (MOIH). The preprohormones of the subfamily-I have a CHH precursor related peptide (CPRP) that is cleaved off during processing; and preprohormones of the subfamily-II lack the CPRP (9). Uncovering their physiological functions will provide new insights into neuroendocrine regulation of energy homeostasis.Characterization of CHH-family neuropeptides is challenging. They are comprised of more than 70 amino acids and often contain multiple post-translational modifications (PTMs) and complex disulfide bridge connections (7). In addition, physiological concentrations of these peptide hormones are typically below picomolar level, and most crustacean species do not have available genome and proteome databases to assist MS-based sequencing.MS-based neuropeptidomics provides a powerful tool for rapid discovery and analysis of a large number of endogenous peptides from the brain and the central nervous system. Our group and others have greatly expanded the peptidomes of many model organisms (3, 18–33). For example, we have discovered more than 200 neuropeptides with several neuropeptide families consisting of as many as 20–40 members in a simple crustacean model system (5, 6, 25–31, 34). However, a majority of these neuropeptides are small peptides with 5–15 amino acid residues long, leaving a gap of identifying larger signaling peptides from organisms without sequenced genome. The observed lack of larger size peptide hormones can be attributed to the lack of effective de novo sequencing strategies for neuropeptides larger than 4 kDa, which are inherently more difficult to fragment using conventional techniques (34–37). Although classical proteomics studies examine larger proteins, these tools are limited to identification based on database searching with one or more peptides matching without complete amino acid sequence coverage (36, 38).Large populations of neuropeptides from 4–10 kDa exist in the nervous systems of both vertebrates and invertebrates (9, 39, 40). Understanding their functional roles requires sufficient molecular knowledge and a unique analytical approach. Therefore, developing effective and reliable methods for de novo sequencing of large neuropeptides at the individual amino acid residue level is an urgent gap to fill in neurobiology. In this study, we present a multifaceted MS strategy aimed at high-definition de novo sequencing and comprehensive characterization of the CHH-family neuropeptides in crustacean central nervous system. The high-definition de novo sequencing was achieved by a combination of three methods: (1) enzymatic digestion and LC-tandem mass spectrometry (MS/MS) bottom-up analysis to generate detailed sequences of proteolytic peptides; (2) off-line LC fractionation and subsequent top-down MS/MS to obtain high-quality fragmentation maps of intact peptides; and (3) on-line LC coupled to top-down MS/MS to allow rapid sequence analysis of low abundance peptides. Combining the three methods overcomes the limitations of each, and thus offers complementary and high-confidence determination of amino acid residues. We report the complete sequence analysis of six CHH-family neuropeptides including the discovery of two novel peptides. With the accurate molecular information, MALDI imaging and ion mobility MS were conducted for the first time to explore their anatomical distribution and biochemical properties.