Martin Steffen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Martin Steffen has a total of 177 co-authors. Here are the top co-authors by number of shared publications:
- Simon Kasif
Bioinformatics Graduate Program, Boston University, Boston, MA, USA
6 shared publications - John Samuelson
Department of Molecular and Cell Biology, Boston University Goldman School of Dental Medicine, Boston, MA 02118, USA
5 shared publications Phillips W. Robbins
Department of Molecular and Cell Biology, Boston University Goldman School of Dental Medicine, Boston, Massachusetts 02118
5 shared publications- Richard John Roberts
New England Biolabs, Ipswich, MA 01938-2723, USA
5 shared publications Brian P. Anton
New England Biolabs, Ipswich, MA 01938-2723, USA
4 shared publicationsDaniel M. Ratner
Department of Molecular and Cell Biology, Boston University Goldman School of Dental Medicine, Boston, Massachusetts 02118
4 shared publications
Latest Publications
- Open Access
Not All Experimental Questions Are Created Equal: Accelerating Biological Data to Knowledge Transformation (BD2K) via Science Informatics, Active Learning and Artificial Intelligence
Pablo Picasso, when first told about computers, famously quipped “Computers are useless. They can only give you answers.” Indeed, the majority of effort in the first half-century of computational research has focused on methods for producing answers. Incredible progress has been achieved in computational modeling, simulation and optimization, across domains as diverse as astrophysics, climate studies, biomedicine, architecture, and chess. However, the use of computers to pose new questions, or prioritize existing ones, has thus far been quite limited. Picasso’s comment highlights the point that good questions can sometimes be more elusive than good answers. The history of science offers numerous examples of the impact of good questions. Paul Erdős, the wandering monk of mathematical graph theory, offered small prizes for anyone who could prove conjectures he identified as important ( 1 ). The prizes varied in cash amounts based on the perceived complexity of the problem posed by Erdős. Posing technical questions and allocating resources to answer them has taken on a new guise in the Internet age. The X-Prize foundation ( http://www.xprize.org/ ) offers multi-million dollar bounties for grand technological goals, including goals for sequencing genomes or space exploration. Several companies provide portals where customers can place cash bounties on educational, scientific or technological challenges, while potential problem solvers can compete to produce the best solutions for these problems. Amazon’s Turk site ( https://www.mturk.com/mturk/welcome ) links people requesting performance of intellectual tasks to people willing to work on them for a fee. Such crowd-sourcing systems create markets of questions and answers, and can help allocate resources and capabilities efficiently. This paradigm suggests a number of interesting questions for scientific research. In a resource limited environment, can funds and research capacity be allocated more efficiently? Can knowledge demand provide an alternative or complementary mechanism to traditional investigator-initiated research grants? The fathers of Artificial Intelligence (AI) and Herbert Simon in particular envisioned the application of AI to Scientific Discovery in different forms and styles (focusing on physics). We follow on these early dreams and describe a novel approach aimed at remodeling of the biomedical research infrastructure and catalyze gene function determination. We aim to start a bold discussion of new ideas aimed towards increasing the efficiency of the allocation of research capacities, reproducibility, provenance tracking, removing redundancy and catalyzing knowledge gain with each experiment. In particular, we describe a tractable computational framework and infrastructure that can help researchers assess the potential information gain of millions of experiments before conducting them. The utility of experiments in this case is modeled as the predictive knowledge (formalized as information) to be gained as a result of performing the experiment. The experimentalist would then be empowered to select experiments that maximized information gain if they wished, recognizing that there are frequently other considerations, such as a specific technological or medical utility, that might over-ride the priority of maximizing information gain. The conceptual approach we develop is general, and here we apply it to the study of gene function.
Open AccessNot All Experimental Questions Are Created Equal: Accelerating Biological Data to Knowledge Transformation (BD2K) via Science Informatics, Active Learning and Artificial Intelligence
Pablo Picasso, when first told about computers, famously quipped “Computers are useless. They can only give you answers.” Indeed, the majority of effort in the first half-century of computational research has focused on methods for producing answers. Incredible progress has been achieved in computational modeling, simulation and optimization, across domains as diverse as astrophysics, climate studies, biomedicine, architecture, and chess. However, the use of computers to pose new questions, or prioritize existing ones, has thus far been quite limited. Picasso’s comment highlights the point that good questions can sometimes be more elusive than good answers. The history of science offers numerous examples of the impact of good questions. Paul Erdős, the wandering monk of mathematical graph theory, offered small prizes for anyone who could prove conjectures he identified as important ( 1 ). The prizes varied in cash amounts based on the perceived complexity of the problem posed by Erdős. Posing technical questions and allocating resources to answer them has taken on a new guise in the Internet age. The X-Prize foundation ( http://www.xprize.org/ ) offers multi-million dollar bounties for grand technological goals, including goals for sequencing genomes or space exploration. Several companies provide portals where customers can place cash bounties on educational, scientific or technological challenges, while potential problem solvers can compete to produce the best solutions for these problems. Amazon’s Turk site ( https://www.mturk.com/mturk/welcome ) links people requesting performance of intellectual tasks to people willing to work on them for a fee. Such crowd-sourcing systems create markets of questions and answers, and can help allocate resources and capabilities efficiently. This paradigm suggests a number of interesting questions for scientific research. In a resource limited environment, can funds and research capacity be allocated more efficiently? Can knowledge demand provide an alternative or complementary mechanism to traditional investigator-initiated research grants? The fathers of Artificial Intelligence (AI) and Herbert Simon in particular envisioned the application of AI to Scientific Discovery in different forms and styles (focusing on physics). We follow on these early dreams and describe a novel approach aimed at remodeling of the biomedical research infrastructure and catalyze gene function determination. We aim to start a bold discussion of new ideas aimed towards increasing the efficiency of the allocation of research capacities, reproducibility, provenance tracking, removing redundancy and catalyzing knowledge gain with each experiment. In particular, we describe a tractable computational framework and infrastructure that can help researchers assess the potential information gain of millions of experiments before conducting them. The utility of experiments in this case is modeled as the predictive knowledge (formalized as information) to be gained as a result of performing the experiment. The experimentalist would then be empowered to select experiments that maximized information gain if they wished, recognizing that there are frequently other considerations, such as a specific technological or medical utility, that might over-ride the priority of maximizing information gain. The conceptual approach we develop is general, and here we apply it to the study of gene function.
- Open Access
- 14 June 2016
- Published by Springer Nature in BMC Molecular Biology
- Vol. 17 (1) , 1-14
Confirmation of translatability and functionality certifies the dual endothelin1/VEGFsp receptor (DEspR) protein
In contrast to rat and mouse databases, the NCBI gene database lists the human dual-endothelin1/VEGFsp receptor (DEspR, formerly Dear) as a unitary transcribed pseudogene due to a stop [TGA]-codon at codon#14 in automated DNA and RNA sequences. However, re-analysis is needed given prior single gene studies detected a tryptophan [TGG]-codon#14 by manual Sanger sequencing, demonstrated DEspR translatability and functionality, and since the demonstration of actual non-translatability through expression studies, the standard-of-excellence for pseudogene designation, has not been performed. Re-analysis must meet UNIPROT criteria for demonstration of a protein’s existence at the highest (protein) level, which a priori, would override DNA- or RNA-based deductions. To dissect the nucleotide sequence discrepancy, we performed Maxam–Gilbert sequencing and reviewed 727 RNA-seq entries. To comply with the highest level multiple UNIPROT criteria for determining DEspR’s existence, we performed various experiments using multiple anti-DEspR monoclonal antibodies (mAbs) targeting distinct DEspR epitopes with one spanning the contested tryptophan [TGG]-codon#14, assessing: (a) DEspR protein expression, (b) predicted full-length protein size, (c) sequence-predicted protein-specific properties beyond codon#14: receptor glycosylation and internalization, (d) protein-partner interactions, and (e) DEspR functionality via DEspR-inhibition effects. Maxam–Gilbert sequencing and some RNA-seq entries demonstrate two guanines, hence a tryptophan [TGG]-codon#14 within a compression site spanning an error-prone compression sequence motif. Western blot analysis using anti-DEspR mAbs targeting distinct DEspR epitopes detect the identical glycosylated 17.5 kDa pull-down protein. Decrease in DEspR-protein size after PNGase-F digest demonstrates post-translational glycosylation, concordant with the consensus-glycosylation site beyond codon#14. Like other small single-transmembrane proteins, mass spectrometry analysis of anti-DEspR mAb pull-down proteins do not detect DEspR, but detect DEspR-protein interactions with proteins implicated in intracellular trafficking and cancer. FACS analyses also detect DEspR-protein in different human cancer stem-like cells (CSCs). DEspR-inhibition studies identify DEspR-roles in CSC survival and growth. Live cell imaging detects fluorescently-labeled anti-DEspR mAb targeted-receptor internalization, concordant with the single internalization-recognition sequence also located beyond codon#14. Data confirm translatability of DEspR, the full-length DEspR protein beyond codon#14, and elucidate DEspR-specific functionality. Along with detection of the tryptophan [TGG]-codon#14 within an error-prone compression site, cumulative data demonstrating DEspR protein existence fulfill multiple UNIPROT criteria, thus refuting its pseudogene designation.
- 14 June 2016
- Published by Springer Nature in BMC Molecular Biology
- Vol. 17 (1) , 1-14
Open AccessConfirmation of translatability and functionality certifies the dual endothelin1/VEGFsp receptor (DEspR) protein
In contrast to rat and mouse databases, the NCBI gene database lists the human dual-endothelin1/VEGFsp receptor (DEspR, formerly Dear) as a unitary transcribed pseudogene due to a stop [TGA]-codon at codon#14 in automated DNA and RNA sequences. However, re-analysis is needed given prior single gene studies detected a tryptophan [TGG]-codon#14 by manual Sanger sequencing, demonstrated DEspR translatability and functionality, and since the demonstration of actual non-translatability through expression studies, the standard-of-excellence for pseudogene designation, has not been performed. Re-analysis must meet UNIPROT criteria for demonstration of a protein’s existence at the highest (protein) level, which a priori, would override DNA- or RNA-based deductions. To dissect the nucleotide sequence discrepancy, we performed Maxam–Gilbert sequencing and reviewed 727 RNA-seq entries. To comply with the highest level multiple UNIPROT criteria for determining DEspR’s existence, we performed various experiments using multiple anti-DEspR monoclonal antibodies (mAbs) targeting distinct DEspR epitopes with one spanning the contested tryptophan [TGG]-codon#14, assessing: (a) DEspR protein expression, (b) predicted full-length protein size, (c) sequence-predicted protein-specific properties beyond codon#14: receptor glycosylation and internalization, (d) protein-partner interactions, and (e) DEspR functionality via DEspR-inhibition effects. Maxam–Gilbert sequencing and some RNA-seq entries demonstrate two guanines, hence a tryptophan [TGG]-codon#14 within a compression site spanning an error-prone compression sequence motif. Western blot analysis using anti-DEspR mAbs targeting distinct DEspR epitopes detect the identical glycosylated 17.5 kDa pull-down protein. Decrease in DEspR-protein size after PNGase-F digest demonstrates post-translational glycosylation, concordant with the consensus-glycosylation site beyond codon#14. Like other small single-transmembrane proteins, mass spectrometry analysis of anti-DEspR mAb pull-down proteins do not detect DEspR, but detect DEspR-protein interactions with proteins implicated in intracellular trafficking and cancer. FACS analyses also detect DEspR-protein in different human cancer stem-like cells (CSCs). DEspR-inhibition studies identify DEspR-roles in CSC survival and growth. Live cell imaging detects fluorescently-labeled anti-DEspR mAb targeted-receptor internalization, concordant with the single internalization-recognition sequence also located beyond codon#14. Data confirm translatability of DEspR, the full-length DEspR protein beyond codon#14, and elucidate DEspR-specific functionality. Along with detection of the tryptophan [TGG]-codon#14 within an error-prone compression site, cumulative data demonstrating DEspR protein existence fulfill multiple UNIPROT criteria, thus refuting its pseudogene designation.
- Open Access
- 3 December 2015
- Published by Oxford University Press (OUP) in Nucleic Acids Research
- Vol. 44 (D1) , D330-D335
COMBREX-DB: an experiment centered database of protein function: knowledge, predictions and knowledge gaps
The COMBREX database (COMBREX-DB; combrex.bu.edu) is an online repository of information related to (i) experimentally determined protein function, (ii) predicted protein function, (iii) relationships among proteins of unknown function and various types of experimental data, including molecular function, protein structure, and associated phenotypes. The database was created as part of the novel COMBREX (COMputational BRidges to EXperiments) effort aimed at accelerating the rate of gene function validation. It currently holds information on ∼3.3 million known and predicted proteins from over 1000 completely sequenced bacterial and archaeal genomes. The database also contains a prototype recommendation system for helping users identify those proteins whose experimental determination of function would be most informative for predicting function for other proteins within protein families. The emphasis on documenting experimental evidence for function predictions, and the prioritization of uncharacterized proteins for experimental testing distinguish COMBREX from other publicly available microbial genomics resources. This article describes updates to COMBREX-DB since an initial description in the 2011 NAR Database Issue.
- 3 December 2015
- Published by Oxford University Press (OUP) in Nucleic Acids Research
- Vol. 44 (D1) , D330-D335
Open AccessCOMBREX-DB: an experiment centered database of protein function: knowledge, predictions and knowledge gaps
The COMBREX database (COMBREX-DB; combrex.bu.edu) is an online repository of information related to (i) experimentally determined protein function, (ii) predicted protein function, (iii) relationships among proteins of unknown function and various types of experimental data, including molecular function, protein structure, and associated phenotypes. The database was created as part of the novel COMBREX (COMputational BRidges to EXperiments) effort aimed at accelerating the rate of gene function validation. It currently holds information on ∼3.3 million known and predicted proteins from over 1000 completely sequenced bacterial and archaeal genomes. The database also contains a prototype recommendation system for helping users identify those proteins whose experimental determination of function would be most informative for predicting function for other proteins within protein families. The emphasis on documenting experimental evidence for function predictions, and the prioritization of uncharacterized proteins for experimental testing distinguish COMBREX from other publicly available microbial genomics resources. This article describes updates to COMBREX-DB since an initial description in the 2011 NAR Database Issue.
- Open Access
- 8 July 2014
- Published by Frontiers Media SA in Frontiers in Genetics
- Vol. 5, 210
Objective: biochemical function
DNA sequencing enables the discovery of new genes in high-throughput, low-cost experiments. Conversely, gene function is determined by low-throughput, high-cost experiments. This inverse relationship for these two types of data is a major impediment in meeting one of the major scientific challenges of our time—the understanding of genomes. This mismatch in throughput is illustrated by considering the progress made for one of the earliest sequenced genomes, that of Mycobacterium tuberculosis H37Rv (Mtb). When its genome was published in 1998, more than a quarter of its genes had no known function (Cole et al., 1998). Our lack of knowledge about these approximately 1000 “conserved hypothetical” genes in Mtb represents a serious deficiency in our understanding of its biology. Now, after more than a decade of progress, our knowledge of those proteins' functions is essentially unchanged—there are still greater than 900 genes with no known function (Lew et al., 2011). In contrast, during this same period, the scientific community has sequenced approximately 18,000 new genomes (Pagani et al., 2012), containing millions of new hypothetical proteins. Apparently, the vector of our progress has tipped decisively away from data interpretation and comprehension, and toward mere data collection. To address the issue of gene function testing and annotation for all microbes, we founded COMBREX (COMputational BRidge to EXperiments), an endeavor aimed at accelerating the rate of gene function validation (Anton et al., 2013). Two of COMBREX's more prominent initiatives were the creation of a comprehensive database for protein function data (http://combrex.bu.edu), and the deployment of a crowdsourcing platform to catalyze protein function experimentation. In the course of these two efforts, it became apparent that fundamental changes in approaches to the problem of protein function determination were needed if there was any hope of keeping pace with DNA sequencing. We suggest that the community work together to (1) re-establish the connection between existing gene annotation and the foundational experimental data that supports all annotation, (2) develop experiment design principles to help guide the identification of maximally informative targets for function validation, (3) invest in the development of higher-throughput approaches for the testing of protein function, and (4) provide an expedited publication pathway for reporting experimental results of gene function, analogous to the reporting of newly sequenced genomes in the journal “Standards in Genomic Sciences.” We recently examined protein function status from greater than 1000 completely sequenced microbes (Anton et al., 2013). For 3.3 million identified genes, we can currently document experimentally determined functions for just 0.4% of the proteins (13,665 proteins). 76% of the proteins are annotated using computational methods, and the remaining 24% of proteins (close to 1 million) have no known, or predicted, functions. Thus, a very small number of experimental data points provides the foundation for an enormously disproportionate number of predicted gene function annotations. (While the total number of experimentally characterized proteins is unknown, we estimate the number to be above 50,000). An unavoidable consequence of the fact that only a small proportion of annotations are based directly on experiment is that predicted functions are often based on weak chains of inference. This can greatly contribute to the proliferation of incorrect annotations. When a newly-discovered gene is annotated based on similarity to a experimentally characterized gene, it then, itself, becomes a source for future annotation. As a result, genes that will be annotated in the future may be annotated based upon genes that are themselves far removed from solid experimental evidence. Compounding confusion, in the vast majority of cases, the original experimental source has not been recorded or preserved. One study estimated that for 37 protein families and 7000 sequences, the overall misannotation rate is roughly 40% (Schnoes et al., 2009), yet the vast majority of annotations are frequently unquestioned by many working scientists. In the first phase of the project, COMBREX awarded funds to 14 labs, and 140 proteins were examined. One of the primary criteria for these applications was prior published work using the proposed enzyme assay. The rationale for this was that experimental efficiency will be greatest, and the costs minimized, in laboratories that already have the reagents, equipment, and expertise necessary to perform the experiments quickly and accurately. Research on many of these proteins has been successfully completed, and results have been published (Chatterjee et al., 2012; Clark et al., 2012; Francis et al., 2012; Phillips et al., 2012; Rodionova et al., 2012; Su et al., 2012; Xu et al., 2012; Choi et al., 2013; Elkin et al., 2013), while research on the others is still in progress. When a protein's function is experimentally determined, it not only affects its own annotation, it changes the probability that other proteins that are close in sequence space have a similar annotation. Thus the potential impact of the experiments COMBREX was able to fund is much larger than simply the proteins tested: the 140 proteins reside in Protein Clusters containing in total more than 3200 proteins, which are therefore quite close in sequence space and likely to have similar functions. At a further distance threshold, there are over 60,000 proteins that have BLAST E-values less than 1e-05. The 140 proteins have eight Pfam-defined domains of unknown function (DUFs), resulting in novel predictive insights for all other proteins containing these DUFs (a total of 1610 in the COMBREX Database). Finally, 37 of these 140 proteins contain a total of 28 unique Pfam-defined domains shared with human proteins, providing functional insights that may impact human health. Several of the COMBREX awards went to labs that had participating undergraduate students, highlighting that the types of experiments COMBREX funds meshes well with the interests and capabilities of undergraduate students eager to participate in original research, and with STEM educational goals of many science departments. As an example, undergraduate students at the University of Virginia were able to successfully determine biochemical activities and enzyme kinetics for three uncharacterized proteins (Elkin et al., 2013). COMBREX hopes to replicate these successes as part of an educational component at numerous undergraduate institutions, in a manner analogous to the Small World Initiative, developed at Yale, which enlists undergraduates in the search for new antibiotics (Barral et al., 2014). When confronted today with the task of annotating a newly discovered hypothetical protein, the use of BLAST quickly and robustly identifies homologous proteins. This sometimes provides clues to potential gene function. However, just as often, one is inundated with matches to other hypothetical proteins that reveal little about possible gene function, and obscures similarities to experimentally characterized proteins. We developed a prototype tool, named COMBLAST, to associate query genes with the various types of experimental evidence and data stored in COMBREX. COMBLAST returns results summarized in a format that concisely captures the functional features of similar proteins. COMBLAST output includes a trace to experimental evidence of function via sequence and domain similarity, to available structural information for related proteins, to association with clinically relevant phenotypes such as antibiotic resistance, and other relevant information. The first application of COMBLAST was deployed in a collaboration led by D. Wood and S. Salzberg (Wood et al., 2012). We analyzed 1474 prokaryotic genome annotations in GenBank and identified 25,394 potential genes that were very likely overlooked during the original annotations. COMBREX was able to provide supporting evidence of their protein-coding nature, and we were able to associate 13,633 of the proteins to published biochemical evidence. Providing explicit links to documented proteins represents one approach for supporting annotations of “missing proteins” (Lane et al., 2014), until comprehensive proteomic surveys confirm their expression (Kim et al., 2014). While an efficient and user-friendly interface to the COMBLAST software is under development, when finally deployed, it will enable any scientist to quickly re-assess the validity of any existing annotation, or to generate hypotheses based solidly on existing experimental evidence. The ability to only perform a small number of experiments places a premium on every attempted experiment, making an important consideration the possible amount of information that will be derived from any one experiment. This “information gain” from the experimental analysis of a given protein is dependent on the number of proteins nearby to it in sequence space, as well as the distances of that protein to previously characterized proteins. In the most simplistic sense, characterization of a judiciously chosen protein generates or improves predictions for many other proteins across many genomes, while characterization of a protein related to few or no other proteins may have a much smaller impact. More formally, for function prediction methods that report probabilities with their predictions (Letovsky and Kasif, 2003), the information gain from an experiment can be quantified as the reduction in the estimated probability of prediction error, summed across all predictions. In COMBREX, we implemented a proof-of-concept prioritization scheme that ranked proteins for experimental testing, which roughly paralleled expected trends in information gain. The “ideal” COMBREX target is a protein close to many other uncharacterized proteins, and relatively far from any protein of known function, but not so far that it would preclude high quality predictions of the protein's function for the experimentalist to test. A second, “soft” guideline was the encouragement to test more than a single protein within a family. Typically, there is only a marginal increase in labor to biochemically test three similar proteins in parallel, when one has procured all the reagents, and created all the buffers for the testing of a single protein, yet the information gain can be significantly increased, as one starts to define boundaries of spaces in which contain proteins with a specific function. Put another way, these design principles do not provide answers—they help experimentalists ask better questions. The functional characterization of hypothetical proteins with only remote sequence homology to known proteins can be challenging, as there may be few clues to guide initial experiments. Several groundbreaking efforts have circumvented this obstacle by deploying technologies that utilize a large diverse set of reagents, or cast their net over a large, complex pool of proteins. Yakunin and coworkers (Kuznetsova et al., 2005; Proudfoot et al., 2008) screen individual proteins for general activity using a set of reagents selected to be generically active (testing for broad functionalities, such as phosphatase, dehydrogenase, protease, etc.), which is then followed by the use of more specific substrates. Cravatt and coworkers (Cravatt et al., 2008; Simon and Cravatt, 2010) have pioneered a complementary approach, “activity-based protein profiling,” enriching enzymes of a particular class using reagents that contain affinity labels, reactive groups and a tag for isolation, and then identifying proteins by mass spectrometry. They and others have applied this technique to multiple classes of enzymes including: hydrolases, proteases, kinases, phosphatases, histone deacetylases, glycosidases, and oxidoreductases. We have recently developed a workflow for the characterization of hypothetical proteins and applied it to six proteins from H. pylori (Choi et al., 2013). We utilized an affinity method to generate initial hypotheses for hypothetical proteins, and then confirmed reactivity using standard recombinant DNA technology and traditional in vitro biochemistry. The affinity reagents utilize nano-particles coated with substrate analogs to enrich proteins from cell lysates of H. pylori. Isolated proteins were identified using mass spectrometry. After cloning and expression in E. coli, the proteins were tested for biochemical activities related to the molecular fragment serving as the affinity bait. Proteins characterized include a guanosine triphosphate (GTP) cyclohydrolase (HP0959), an ATPase (HP1079), an adenosine deaminase (HP0267), a phosphodiesterase (HP1042), an aminopeptidase (HP1037), and new substrates were characterized for a peptidoglycan deacetylase (HP0310). We suspect that a tremendous amount of pertinent experimental gene function information is lost to the community at large because of difficulties associated with finding appropriate venues to disseminate the information. The genomics community addressed this need smartly with the creation of an open access journal, Standards in Genomic Sciences. This journal typically publishes short, straightforward descriptions reporting a new genome sequence based on a standard template. There is a need for a similar publication mechanism for gene function data. It appears that currently, the scientific community's publication standards generally dictate that a successful biochemical experiment alone does not meet the criteria for a minimum publishable unit. Without accompanying data about the gene's role in the biology of the organism, or observations on associated phenotypic effects, biochemical results are not “enough” of a story. As a result, useful experimental information remains hidden in individual notebooks, lost to the wider community. In our opinion, there would be great value in a publication venue that accepted streamlined “biochemical reports” in a routine manner. Minimal data provided would be the sequence of the gene, the protein production method, the biochemical assay, and an interpretation of the results. Similarly, simple reports on gene overexpression or knockouts and their phenotypic effects would permit the dissemination of meaningful functional data. Such data could be linked to COMBREX and other frequently accessed gene databases to expedite the dissemination process by avoiding human curation or processing. There needs to be a paradigm shift in the approach taken to determine and assign gene function if there is to be any hope of realizing the potential benefits from the torrent of new genome sequences. We advocate here for: (1) experimental designs that test sets of maximally informative proteins, (2) maximal information extraction from every experimental result, with explicit traces provided to related proteins, (3) enhanced opportunities for collaboration among computational and experimental researchers to share predictions and results, and distribute limited resources, (4) investment by granting agencies in the development of high-throughput gene function testing, and (5) the creation of new publication options to report and share the results of experiments that are performed. COMBREX is funded by a GO grant from the National Institute of General Medical Sciences (NIGMS) (1RC2GM092602-01). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Anton, B. P., Chang, Y.-C., Brown, P., Choi, H.-P., Faller, L. L., Guleria, J., et al. (2013). The COMBREX project: design, methodology, and initial results. PLoS Biol. 11:e1001638. doi: 10.1371/journal.pbio.1001638 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Barral, A. M., Makhluf, H., Soneral, P., and Gasper, B. (2014). Small World Initiative: crowdsourcing research of new antibiotics to enhance undergraduate biology teaching (618.41). FASEB J. 28, 618–641. Chatterjee, K., Blaby, I. K., Thiaville, P. C., Majumder, M., Grosjean, H., Yuan, Y. A., et al. (2012). The archaeal COG1901/DUF358 SPOUT-methyltransferase members, together with pseudouridine synthase Pus10, catalyze the formation of 1-methylpseudouridine at position 54 of tRNA. RNA 18, 421–433. doi: 10.1261/rna.030841.111 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Choi, H. P., Juarez, S., Ciordia, S., Fernandez, M., Bargiela, R., Albar, J. P., et al. (2013). Biochemical characterization of hypothetical proteins from Helicobacter pylori. PLoS ONE 8:e66605. doi: 10.1371/journal.pone.0066605 CrossRef Full Text Clark, T. A., Murray, I. A., Morgan, R. D., Kislyuk, A. O., Spittle, K. E., Boitano, M., et al. (2012). Characterization of DNA methyltransferase specificities using single-molecule, real-time DNA sequencing. Nucleic Acids Res. 40:e29. doi: 10.1093/nar/gkr1146 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Cole, S. T., Brosch, R., Parkhill, J., Garnier, T., Churcher, C., Harris, D., et al. (1998). Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature 393, 537–544. doi: 10.1038/31159 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Cravatt, B. F., Wright, A. T., and Kozarich, J. W. (2008). Activity-based protein profiling: from enzyme chemistry to proteomic chemistry. Annu. Rev. Biochem. 77, 383–414. doi: 10.1146/annurev.biochem.75.101304.124125 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Elkin, S. R., Kumar, A., Price, C. W., and Columbus, L. (2013). A broad specificity nucleoside kinase from Thermoplasma acidophilum. Proteins 81, 568–582. doi: 10.1002/prot.24212 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Francis, K., Nishino, S. F., Spain, J. C., and Gadda, G. (2012). A novel activity for fungal nitronate monooxygenase: detoxification of the metabolic inhibitor propionate-3-nitronate. Arch. Biochem. Biophys. 521, 84–89. doi: 10.1016/j.abb.2012.03.015 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Kim, M. S., Pinto, S. M., Getnet, D., Nirujogi, R. S., Manda, S. S., Chaerkady, R., et al. (2014). A draft map of the human proteome. Nature 509, 575–581. doi: 10.1038/nature13302 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Kuznetsova, E., Proudfoot, M., Sanders, S. A., Reinking, J., Savchenko, A., Arrowsmith, C. H., et al. (2005). Enzyme genomics: application of general enzymatic screens to discover new enzymes. FEMS Microbiol. Rev. 29, 263–279. doi: 10.1016/j.femsre.2004.12.006 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Lane, L., Bairoch, A., Beavis, R. C., Deutsch, E. W., Gaudet, P., Lundberg, E., et al. (2014). Metrics for the Human Proteome Project 2013-2014 and strategies for finding missing proteins. J. Proteome Res. 13, 15–20. doi: 10.1021/pr401144x Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Letovsky, S., and Kasif, S. (2003). Predicting protein function from protein/protein interaction data: a probabilistic approach. Bioinformatics 19(Suppl. 1), i197–i204. doi: 10.1093/bioinformatics/btg1026 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Lew, J. M., Kapopoulou, A., Jones, L. M., and Cole, S. T. (2011). TubercuList–10 years after. Tuberculosis (Edinb.) 91, 1–7. doi: 10.1016/j.tube.2010.09.008 Pubmed Abstract | Pubmed Full Text | CrossRef...
- 8 July 2014
- Published by Frontiers Media SA in Frontiers in Genetics
- Vol. 5, 210
Open AccessObjective: biochemical function
DNA sequencing enables the discovery of new genes in high-throughput, low-cost experiments. Conversely, gene function is determined by low-throughput, high-cost experiments. This inverse relationship for these two types of data is a major impediment in meeting one of the major scientific challenges of our time—the understanding of genomes. This mismatch in throughput is illustrated by considering the progress made for one of the earliest sequenced genomes, that of Mycobacterium tuberculosis H37Rv (Mtb). When its genome was published in 1998, more than a quarter of its genes had no known function (Cole et al., 1998). Our lack of knowledge about these approximately 1000 “conserved hypothetical” genes in Mtb represents a serious deficiency in our understanding of its biology. Now, after more than a decade of progress, our knowledge of those proteins' functions is essentially unchanged—there are still greater than 900 genes with no known function (Lew et al., 2011). In contrast, during this same period, the scientific community has sequenced approximately 18,000 new genomes (Pagani et al., 2012), containing millions of new hypothetical proteins. Apparently, the vector of our progress has tipped decisively away from data interpretation and comprehension, and toward mere data collection. To address the issue of gene function testing and annotation for all microbes, we founded COMBREX (COMputational BRidge to EXperiments), an endeavor aimed at accelerating the rate of gene function validation (Anton et al., 2013). Two of COMBREX's more prominent initiatives were the creation of a comprehensive database for protein function data (http://combrex.bu.edu), and the deployment of a crowdsourcing platform to catalyze protein function experimentation. In the course of these two efforts, it became apparent that fundamental changes in approaches to the problem of protein function determination were needed if there was any hope of keeping pace with DNA sequencing. We suggest that the community work together to (1) re-establish the connection between existing gene annotation and the foundational experimental data that supports all annotation, (2) develop experiment design principles to help guide the identification of maximally informative targets for function validation, (3) invest in the development of higher-throughput approaches for the testing of protein function, and (4) provide an expedited publication pathway for reporting experimental results of gene function, analogous to the reporting of newly sequenced genomes in the journal “Standards in Genomic Sciences.” We recently examined protein function status from greater than 1000 completely sequenced microbes (Anton et al., 2013). For 3.3 million identified genes, we can currently document experimentally determined functions for just 0.4% of the proteins (13,665 proteins). 76% of the proteins are annotated using computational methods, and the remaining 24% of proteins (close to 1 million) have no known, or predicted, functions. Thus, a very small number of experimental data points provides the foundation for an enormously disproportionate number of predicted gene function annotations. (While the total number of experimentally characterized proteins is unknown, we estimate the number to be above 50,000). An unavoidable consequence of the fact that only a small proportion of annotations are based directly on experiment is that predicted functions are often based on weak chains of inference. This can greatly contribute to the proliferation of incorrect annotations. When a newly-discovered gene is annotated based on similarity to a experimentally characterized gene, it then, itself, becomes a source for future annotation. As a result, genes that will be annotated in the future may be annotated based upon genes that are themselves far removed from solid experimental evidence. Compounding confusion, in the vast majority of cases, the original experimental source has not been recorded or preserved. One study estimated that for 37 protein families and 7000 sequences, the overall misannotation rate is roughly 40% (Schnoes et al., 2009), yet the vast majority of annotations are frequently unquestioned by many working scientists. In the first phase of the project, COMBREX awarded funds to 14 labs, and 140 proteins were examined. One of the primary criteria for these applications was prior published work using the proposed enzyme assay. The rationale for this was that experimental efficiency will be greatest, and the costs minimized, in laboratories that already have the reagents, equipment, and expertise necessary to perform the experiments quickly and accurately. Research on many of these proteins has been successfully completed, and results have been published (Chatterjee et al., 2012; Clark et al., 2012; Francis et al., 2012; Phillips et al., 2012; Rodionova et al., 2012; Su et al., 2012; Xu et al., 2012; Choi et al., 2013; Elkin et al., 2013), while research on the others is still in progress. When a protein's function is experimentally determined, it not only affects its own annotation, it changes the probability that other proteins that are close in sequence space have a similar annotation. Thus the potential impact of the experiments COMBREX was able to fund is much larger than simply the proteins tested: the 140 proteins reside in Protein Clusters containing in total more than 3200 proteins, which are therefore quite close in sequence space and likely to have similar functions. At a further distance threshold, there are over 60,000 proteins that have BLAST E-values less than 1e-05. The 140 proteins have eight Pfam-defined domains of unknown function (DUFs), resulting in novel predictive insights for all other proteins containing these DUFs (a total of 1610 in the COMBREX Database). Finally, 37 of these 140 proteins contain a total of 28 unique Pfam-defined domains shared with human proteins, providing functional insights that may impact human health. Several of the COMBREX awards went to labs that had participating undergraduate students, highlighting that the types of experiments COMBREX funds meshes well with the interests and capabilities of undergraduate students eager to participate in original research, and with STEM educational goals of many science departments. As an example, undergraduate students at the University of Virginia were able to successfully determine biochemical activities and enzyme kinetics for three uncharacterized proteins (Elkin et al., 2013). COMBREX hopes to replicate these successes as part of an educational component at numerous undergraduate institutions, in a manner analogous to the Small World Initiative, developed at Yale, which enlists undergraduates in the search for new antibiotics (Barral et al., 2014). When confronted today with the task of annotating a newly discovered hypothetical protein, the use of BLAST quickly and robustly identifies homologous proteins. This sometimes provides clues to potential gene function. However, just as often, one is inundated with matches to other hypothetical proteins that reveal little about possible gene function, and obscures similarities to experimentally characterized proteins. We developed a prototype tool, named COMBLAST, to associate query genes with the various types of experimental evidence and data stored in COMBREX. COMBLAST returns results summarized in a format that concisely captures the functional features of similar proteins. COMBLAST output includes a trace to experimental evidence of function via sequence and domain similarity, to available structural information for related proteins, to association with clinically relevant phenotypes such as antibiotic resistance, and other relevant information. The first application of COMBLAST was deployed in a collaboration led by D. Wood and S. Salzberg (Wood et al., 2012). We analyzed 1474 prokaryotic genome annotations in GenBank and identified 25,394 potential genes that were very likely overlooked during the original annotations. COMBREX was able to provide supporting evidence of their protein-coding nature, and we were able to associate 13,633 of the proteins to published biochemical evidence. Providing explicit links to documented proteins represents one approach for supporting annotations of “missing proteins” (Lane et al., 2014), until comprehensive proteomic surveys confirm their expression (Kim et al., 2014). While an efficient and user-friendly interface to the COMBLAST software is under development, when finally deployed, it will enable any scientist to quickly re-assess the validity of any existing annotation, or to generate hypotheses based solidly on existing experimental evidence. The ability to only perform a small number of experiments places a premium on every attempted experiment, making an important consideration the possible amount of information that will be derived from any one experiment. This “information gain” from the experimental analysis of a given protein is dependent on the number of proteins nearby to it in sequence space, as well as the distances of that protein to previously characterized proteins. In the most simplistic sense, characterization of a judiciously chosen protein generates or improves predictions for many other proteins across many genomes, while characterization of a protein related to few or no other proteins may have a much smaller impact. More formally, for function prediction methods that report probabilities with their predictions (Letovsky and Kasif, 2003), the information gain from an experiment can be quantified as the reduction in the estimated probability of prediction error, summed across all predictions. In COMBREX, we implemented a proof-of-concept prioritization scheme that ranked proteins for experimental testing, which roughly paralleled expected trends in information gain. The “ideal” COMBREX target is a protein close to many other uncharacterized proteins, and relatively far from any protein of known function, but not so far that it would preclude high quality predictions of the protein's function for the experimentalist to test. A second, “soft” guideline was the encouragement to test more than a single protein within a family. Typically, there is only a marginal increase in labor to biochemically test three similar proteins in parallel, when one has procured all the reagents, and created all the buffers for the testing of a single protein, yet the information gain can be significantly increased, as one starts to define boundaries of spaces in which contain proteins with a specific function. Put another way, these design principles do not provide answers—they help experimentalists ask better questions. The functional characterization of hypothetical proteins with only remote sequence homology to known proteins can be challenging, as there may be few clues to guide initial experiments. Several groundbreaking efforts have circumvented this obstacle by deploying technologies that utilize a large diverse set of reagents, or cast their net over a large, complex pool of proteins. Yakunin and coworkers (Kuznetsova et al., 2005; Proudfoot et al., 2008) screen individual proteins for general activity using a set of reagents selected to be generically active (testing for broad functionalities, such as phosphatase, dehydrogenase, protease, etc.), which is then followed by the use of more specific substrates. Cravatt and coworkers (Cravatt et al., 2008; Simon and Cravatt, 2010) have pioneered a complementary approach, “activity-based protein profiling,” enriching enzymes of a particular class using reagents that contain affinity labels, reactive groups and a tag for isolation, and then identifying proteins by mass spectrometry. They and others have applied this technique to multiple classes of enzymes including: hydrolases, proteases, kinases, phosphatases, histone deacetylases, glycosidases, and oxidoreductases. We have recently developed a workflow for the characterization of hypothetical proteins and applied it to six proteins from H. pylori (Choi et al., 2013). We utilized an affinity method to generate initial hypotheses for hypothetical proteins, and then confirmed reactivity using standard recombinant DNA technology and traditional in vitro biochemistry. The affinity reagents utilize nano-particles coated with substrate analogs to enrich proteins from cell lysates of H. pylori. Isolated proteins were identified using mass spectrometry. After cloning and expression in E. coli, the proteins were tested for biochemical activities related to the molecular fragment serving as the affinity bait. Proteins characterized include a guanosine triphosphate (GTP) cyclohydrolase (HP0959), an ATPase (HP1079), an adenosine deaminase (HP0267), a phosphodiesterase (HP1042), an aminopeptidase (HP1037), and new substrates were characterized for a peptidoglycan deacetylase (HP0310). We suspect that a tremendous amount of pertinent experimental gene function information is lost to the community at large because of difficulties associated with finding appropriate venues to disseminate the information. The genomics community addressed this need smartly with the creation of an open access journal, Standards in Genomic Sciences. This journal typically publishes short, straightforward descriptions reporting a new genome sequence based on a standard template. There is a need for a similar publication mechanism for gene function data. It appears that currently, the scientific community's publication standards generally dictate that a successful biochemical experiment alone does not meet the criteria for a minimum publishable unit. Without accompanying data about the gene's role in the biology of the organism, or observations on associated phenotypic effects, biochemical results are not “enough” of a story. As a result, useful experimental information remains hidden in individual notebooks, lost to the wider community. In our opinion, there would be great value in a publication venue that accepted streamlined “biochemical reports” in a routine manner. Minimal data provided would be the sequence of the gene, the protein production method, the biochemical assay, and an interpretation of the results. Similarly, simple reports on gene overexpression or knockouts and their phenotypic effects would permit the dissemination of meaningful functional data. Such data could be linked to COMBREX and other frequently accessed gene databases to expedite the dissemination process by avoiding human curation or processing. There needs to be a paradigm shift in the approach taken to determine and assign gene function if there is to be any hope of realizing the potential benefits from the torrent of new genome sequences. We advocate here for: (1) experimental designs that test sets of maximally informative proteins, (2) maximal information extraction from every experimental result, with explicit traces provided to related proteins, (3) enhanced opportunities for collaboration among computational and experimental researchers to share predictions and results, and distribute limited resources, (4) investment by granting agencies in the development of high-throughput gene function testing, and (5) the creation of new publication options to report and share the results of experiments that are performed. COMBREX is funded by a GO grant from the National Institute of General Medical Sciences (NIGMS) (1RC2GM092602-01). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Anton, B. P., Chang, Y.-C., Brown, P., Choi, H.-P., Faller, L. L., Guleria, J., et al. (2013). The COMBREX project: design, methodology, and initial results. PLoS Biol. 11:e1001638. doi: 10.1371/journal.pbio.1001638 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Barral, A. M., Makhluf, H., Soneral, P., and Gasper, B. (2014). Small World Initiative: crowdsourcing research of new antibiotics to enhance undergraduate biology teaching (618.41). FASEB J. 28, 618–641. Chatterjee, K., Blaby, I. K., Thiaville, P. C., Majumder, M., Grosjean, H., Yuan, Y. A., et al. (2012). The archaeal COG1901/DUF358 SPOUT-methyltransferase members, together with pseudouridine synthase Pus10, catalyze the formation of 1-methylpseudouridine at position 54 of tRNA. RNA 18, 421–433. doi: 10.1261/rna.030841.111 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Choi, H. P., Juarez, S., Ciordia, S., Fernandez, M., Bargiela, R., Albar, J. P., et al. (2013). Biochemical characterization of hypothetical proteins from Helicobacter pylori. PLoS ONE 8:e66605. doi: 10.1371/journal.pone.0066605 CrossRef Full Text Clark, T. A., Murray, I. A., Morgan, R. D., Kislyuk, A. O., Spittle, K. E., Boitano, M., et al. (2012). Characterization of DNA methyltransferase specificities using single-molecule, real-time DNA sequencing. Nucleic Acids Res. 40:e29. doi: 10.1093/nar/gkr1146 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Cole, S. T., Brosch, R., Parkhill, J., Garnier, T., Churcher, C., Harris, D., et al. (1998). Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature 393, 537–544. doi: 10.1038/31159 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Cravatt, B. F., Wright, A. T., and Kozarich, J. W. (2008). Activity-based protein profiling: from enzyme chemistry to proteomic chemistry. Annu. Rev. Biochem. 77, 383–414. doi: 10.1146/annurev.biochem.75.101304.124125 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Elkin, S. R., Kumar, A., Price, C. W., and Columbus, L. (2013). A broad specificity nucleoside kinase from Thermoplasma acidophilum. Proteins 81, 568–582. doi: 10.1002/prot.24212 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Francis, K., Nishino, S. F., Spain, J. C., and Gadda, G. (2012). A novel activity for fungal nitronate monooxygenase: detoxification of the metabolic inhibitor propionate-3-nitronate. Arch. Biochem. Biophys. 521, 84–89. doi: 10.1016/j.abb.2012.03.015 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Kim, M. S., Pinto, S. M., Getnet, D., Nirujogi, R. S., Manda, S. S., Chaerkady, R., et al. (2014). A draft map of the human proteome. Nature 509, 575–581. doi: 10.1038/nature13302 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Kuznetsova, E., Proudfoot, M., Sanders, S. A., Reinking, J., Savchenko, A., Arrowsmith, C. H., et al. (2005). Enzyme genomics: application of general enzymatic screens to discover new enzymes. FEMS Microbiol. Rev. 29, 263–279. doi: 10.1016/j.femsre.2004.12.006 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Lane, L., Bairoch, A., Beavis, R. C., Deutsch, E. W., Gaudet, P., Lundberg, E., et al. (2014). Metrics for the Human Proteome Project 2013-2014 and strategies for finding missing proteins. J. Proteome Res. 13, 15–20. doi: 10.1021/pr401144x Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Letovsky, S., and Kasif, S. (2003). Predicting protein function from protein/protein interaction data: a probabilistic approach. Bioinformatics 19(Suppl. 1), i197–i204. doi: 10.1093/bioinformatics/btg1026 Pubmed Abstract | Pubmed Full Text | CrossRef Full Text Lew, J. M., Kapopoulou, A., Jones, L. M., and Cole, S. T. (2011). TubercuList–10 years after. Tuberculosis (Edinb.) 91, 1–7. doi: 10.1016/j.tube.2010.09.008 Pubmed Abstract | Pubmed Full Text | CrossRef...
- Open Access
- 27 August 2013
- Published by Public Library of Science (PLoS) in PLOS Biology
- Vol. 11 (8) , e1001638
The COMBREX Project: Design, Methodology, and Initial Results
Experimental data exists for only a vanishingly small fraction of sequenced microbial genes. This community page discusses the progress made by the COMBREX project to address this important issue using both computational and experimental resources.
- 27 August 2013
- Published by Public Library of Science (PLoS) in PLOS Biology
- Vol. 11 (8) , e1001638
Open AccessThe COMBREX Project: Design, Methodology, and Initial Results
Experimental data exists for only a vanishingly small fraction of sequenced microbial genes. This community page discusses the progress made by the COMBREX project to address this important issue using both computational and experimental resources.
- Open Access
- 18 June 2013
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 8 (6) , e66605
Biochemical Characterization of Hypothetical Proteins from Helicobacter pylori
The functional characterization of Open Reading Frames (ORFs) from sequenced genomes remains a bottleneck in our effort to understand microbial biology. In particular, the functional characterization of proteins with only remote sequence homology to known proteins can be challenging, as there may be few clues to guide initial experiments. Affinity enrichment of proteins from cell lysates, and a global perspective of protein function as provided by COMBREX, affords an approach to this problem. We present here the biochemical analysis of six proteins from Helicobacter pylori ATCC 26695, a focus organism in COMBREX. Initial hypotheses were based upon affinity capture of proteins from total cellular lysate using derivatized nano-particles, and subsequent identification by mass spectrometry. Candidate genes encoding these proteins were cloned and expressed in Escherichia coli, and the recombinant proteins were purified and characterized biochemically and their biochemical parameters compared with the native ones. These proteins include a guanosine triphosphate (GTP) cyclohydrolase (HP0959), an ATPase (HP1079), an adenosine deaminase (HP0267), a phosphodiesterase (HP1042), an aminopeptidase (HP1037), and new substrates were characterized for a peptidoglycan deacetylase (HP0310). Generally, characterized enzymes were active at acidic to neutral pH (4.0–7.5) with temperature optima ranging from 35 to 55°C, although some exhibited outstanding characteristics.
- 18 June 2013
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 8 (6) , e66605
Open AccessBiochemical Characterization of Hypothetical Proteins from Helicobacter pylori
The functional characterization of Open Reading Frames (ORFs) from sequenced genomes remains a bottleneck in our effort to understand microbial biology. In particular, the functional characterization of proteins with only remote sequence homology to known proteins can be challenging, as there may be few clues to guide initial experiments. Affinity enrichment of proteins from cell lysates, and a global perspective of protein function as provided by COMBREX, affords an approach to this problem. We present here the biochemical analysis of six proteins from Helicobacter pylori ATCC 26695, a focus organism in COMBREX. Initial hypotheses were based upon affinity capture of proteins from total cellular lysate using derivatized nano-particles, and subsequent identification by mass spectrometry. Candidate genes encoding these proteins were cloned and expressed in Escherichia coli, and the recombinant proteins were purified and characterized biochemically and their biochemical parameters compared with the native ones. These proteins include a guanosine triphosphate (GTP) cyclohydrolase (HP0959), an ATPase (HP1079), an adenosine deaminase (HP0267), a phosphodiesterase (HP1042), an aminopeptidase (HP1037), and new substrates were characterized for a peptidoglycan deacetylase (HP0310). Generally, characterized enzymes were active at acidic to neutral pH (4.0–7.5) with temperature optima ranging from 35 to 55°C, although some exhibited outstanding characteristics.
- 1 January 2013
- Published by Elsevier in New Biotechnology
- Vol. 30 (2) , 153-158
Multiplex protein detection with DNA readout via mass spectrometry
- 1 January 2013
- Published by Elsevier in New Biotechnology
- Vol. 30 (2) , 153-158
Multiplex protein detection with DNA readout via mass spectrometry
- 1 September 2012
- Published by Defense Technical Information Center (DTIC)
Proteomics of Peripheral Leukocytes in Patients with Elevated Serum Levels of PSA
During this third year of research, we have analyzed protein abundance levels of CD4+ T-lymphocytes and demonstrated that in accordance with our original hypothesis, we are able to discriminate between prostate cancer and benign prostate disease in patients with an elevated PSA level. A manuscript detailing these results is being prepared. Additionally, we explored the proteomes of lymphocytes from mice, and have found that the procedures involved with sample procurement are less robust than for the human samples, encouraging us to focus on the human samples. We remain optimistic that we could potentially lay the groundwork for a test that would significantly reduce the number of prostate biopsies performed annually.
- 1 September 2012
- Published by Defense Technical Information Center (DTIC)
Proteomics of Peripheral Leukocytes in Patients with Elevated Serum Levels of PSA
During this third year of research, we have analyzed protein abundance levels of CD4+ T-lymphocytes and demonstrated that in accordance with our original hypothesis, we are able to discriminate between prostate cancer and benign prostate disease in patients with an elevated PSA level. A manuscript detailing these results is being prepared. Additionally, we explored the proteomes of lymphocytes from mice, and have found that the procedures involved with sample procurement are less robust than for the human samples, encouraging us to focus on the human samples. We remain optimistic that we could potentially lay the groundwork for a test that would significantly reduce the number of prostate biopsies performed annually.
- Open Access
- 1 January 2012
- Published by Springer Nature in Biology Direct
- Vol. 7 (1) , 37
Thousands of missed genes found in bacterial genomes and their analysis with COMBREX
The dramatic reduction in the cost of sequencing has allowed many researchers to join in the effort of sequencing and annotating prokaryotic genomes. Annotation methods vary considerably and may fail to identify some genes. Here we draw attention to a large number of likely genes missing from annotations using common tools such as Glimmer and BLAST.
- 1 January 2012
- Published by Springer Nature in Biology Direct
- Vol. 7 (1) , 37
Open AccessThousands of missed genes found in bacterial genomes and their analysis with COMBREX
The dramatic reduction in the cost of sequencing has allowed many researchers to join in the effort of sequencing and annotating prokaryotic genomes. Annotation methods vary considerably and may fail to identify some genes. Here we draw attention to a large number of likely genes missing from annotations using common tools such as Glimmer and BLAST.
- Open Access
- 14 April 2011
- Published by Springer Nature in Molecular Medicine
- Vol. 17 (7-8) , 588-598
Prestroke Proteomic Changes in Cerebral Microvessels in Stroke-Prone, Transgenic[hCETP]-Hyperlipidemic, Dahl Salt-Sensitive Hypertensive Rats
Stroke is the third leading cause of death in the United States with high rates of morbidity among survivors. The search to fill the unequivocal need for new therapeutic approaches would benefit from unbiased proteomic analyses of animal models of spontaneous stroke in the prestroke stage. Since brain microvessels play key roles in neurovascular coupling, we investigated prestroke microvascular proteome changes. Proteomic analysis of cerebral cortical microvessels (cMVs) was done by tandem mass spectrometry comparing two prestroke time points. Metaprotein-pathway analyses of proteomic spectral count data were done to identify risk factor-induced changes, followed by QSPEC-analyses of individual protein changes associated with increased stroke susceptibility. We report 26 cMV proteome profiles from male and female stroke-prone and non-stroke-prone rats at 2 months and 4.5 months of age prior to overt stroke events. We identified 1,934 proteins by two or more peptides. Metaprotein pathway analysis detected age-associated changes in energy metabolism and cell-to-microenvironment interactions, as well as sex-specific changes in energy metabolism and endothelial leukocyte transmigration pathways. Stroke susceptibility was associated independently with multiple protein changes associated with ischemia, angiogenesis or involved in blood brain barrier (BBB) integrity. Immunohistochemical analysis confirmed aquaporin-4 and laminin-α1 induction in cMVs, representative of proteomic changes with >65 Bayes factor (BF), associated with stroke susceptibility. Altogether, proteomic analysis demonstrates significant molecular changes in ischemic cerebral microvasculature in the prestroke stage, which could contribute to the observed model phenotype of microhemorrhages and postischemic hemorrhagic transformation. These pathways comprise putative targets for translational research of much needed novel diagnostic and therapeutic approaches for stroke.
- 14 April 2011
- Published by Springer Nature in Molecular Medicine
- Vol. 17 (7-8) , 588-598
Open AccessPrestroke Proteomic Changes in Cerebral Microvessels in Stroke-Prone, Transgenic[hCETP]-Hyperlipidemic, Dahl Salt-Sensitive Hypertensive Rats
Stroke is the third leading cause of death in the United States with high rates of morbidity among survivors. The search to fill the unequivocal need for new therapeutic approaches would benefit from unbiased proteomic analyses of animal models of spontaneous stroke in the prestroke stage. Since brain microvessels play key roles in neurovascular coupling, we investigated prestroke microvascular proteome changes. Proteomic analysis of cerebral cortical microvessels (cMVs) was done by tandem mass spectrometry comparing two prestroke time points. Metaprotein-pathway analyses of proteomic spectral count data were done to identify risk factor-induced changes, followed by QSPEC-analyses of individual protein changes associated with increased stroke susceptibility. We report 26 cMV proteome profiles from male and female stroke-prone and non-stroke-prone rats at 2 months and 4.5 months of age prior to overt stroke events. We identified 1,934 proteins by two or more peptides. Metaprotein pathway analysis detected age-associated changes in energy metabolism and cell-to-microenvironment interactions, as well as sex-specific changes in energy metabolism and endothelial leukocyte transmigration pathways. Stroke susceptibility was associated independently with multiple protein changes associated with ischemia, angiogenesis or involved in blood brain barrier (BBB) integrity. Immunohistochemical analysis confirmed aquaporin-4 and laminin-α1 induction in cMVs, representative of proteomic changes with >65 Bayes factor (BF), associated with stroke susceptibility. Altogether, proteomic analysis demonstrates significant molecular changes in ischemic cerebral microvasculature in the prestroke stage, which could contribute to the observed model phenotype of microhemorrhages and postischemic hemorrhagic transformation. These pathways comprise putative targets for translational research of much needed novel diagnostic and therapeutic approaches for stroke.