Dataset Information

The challenge of increasing Pfam coverage of the human proteome.

ABSTRACT: It is a worthy goal to completely characterize all human proteins in terms of their domains. Here, using the Pfam database, we asked how far we have progressed in this endeavour. Ninety per cent of proteins in the human proteome matched at least one of 5494 manually curated Pfam-A families. In contrast, human residue coverage by Pfam-A families was <45%, with 9418 automatically generated Pfam-B families adding a further 10%. Even after excluding predicted signal peptide regions and short regions (<50 consecutive residues) unlikely to harbour new families, for ?38% of the human protein residues, there was no information in Pfam about conservation and evolutionary relationship with other protein regions. This uncovered portion of the human proteome was found to be distributed over almost 25 000 distinct protein regions. Comparison with proteins in the UniProtKB database suggested that the human regions that exhibited similarity to thousands of other sequences were often either divergent elements or N- or C-terminal extensions of existing families. Thirty-four per cent of regions, on the other hand, matched fewer than 100 sequences in UniProtKB. Most of these did not appear to share any relationship with existing Pfam-A families, suggesting that thousands of new families would need to be generated to cover them. Also, these latter regions were particularly rich in amino acid compositional bias such as the one associated with intrinsic disorder. This could represent a significant obstacle toward their inclusion into new Pfam families. Based on these observations, a major focus for increasing Pfam coverage of the human proteome will be to improve the definition of existing families. New families will also be built, prioritizing those that have been experimentally functionally characterized. Database URL: http://pfam.sanger.ac.uk/

SUBMITTER: Mistry J

PROVIDER: S-EPMC3630804 | biostudies-literature | 2013

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

The challenge of increasing Pfam coverage of the human proteome.

Mistry Jaina J Coggill Penny P Eberhardt Ruth Y RY Deiana Antonio A Giansanti Andrea A Finn Robert D RD Bateman Alex A Punta Marco M

Database : the journal of biological databases and curation 20130419

It is a worthy goal to completely characterize all human proteins in terms of their domains. Here, using the Pfam database, we asked how far we have progressed in this endeavour. Ninety per cent of proteins in the human proteome matched at least one of 5494 manually curated Pfam-A families. In contrast, human residue coverage by Pfam-A families was <45%, with 9418 automatically generated Pfam-B families adding a further 10%. Even after excluding predicted signal peptide regions and short regions ...[more]

PMID: 23603847

Similar Datasets

Project description:Cerebrospinal fluid (CSF) is in direct contact with the central nervous system. This makes human CSF an attractive source of potential biomarkers for neurologic diseases. Similarly to blood plasma, proteomic analysis of CSF is complicated by a high dynamic range of individual protein concentrations and by the presence of several highly abundant proteins. To deal with the abundant human CSF proteins, methods developed for blood plasma/serum are routinely used. Multiple affinity removal systems and protein enrichment of less abundant proteins using a combinatorial peptide ligand library are among the most frequent approaches. However, their relative impact on CSF proteome coverage has never been evaluated side-by-side in a single study. Therefore, we explored the effect of CSF depletion using MARS 14 cartridge and ProteoMiner ligand library on the number of CSF proteins identified in subsequent LC-MS/MS analysis. LC-MS/MS analysis of crude (non-treated) CSF provided roughly 500 identified proteins. Depletion of CSF by MARS 14 cartridge increased the number of identifications to nearly 800, while treatment of CSF using ProteoMiner enabled identification of 600 proteins. To explore the potential losses of CSF proteins during the depletion process, we also analyzed the "waste" fractions generated by both methods, i.e., proteins retained by the MARS 14 cartridge, and the molecules present in the flow-through fraction from ProteoMiner. More than 250 proteins were bound to MARS 14 cartridge, 100 of those were not identified in the corresponding depleted CSF. Similarly, analysis of the waste fraction in ProteoMiner workflow provided almost 70 unique proteins not found in the CSF depleted by the ligand library. Both depletion strategies significantly increased the number of identified CSF proteins compared to crude CSF. However, MARS 14 depletion provided a markedly higher number of identified proteins (773) compared to ProteoMiner (611). Further, we showed that CSF proteins are lost due to co-depletion (MARS 14) or exclusion (ProteoMiner) during the depletion process. This suggests that the routinely discarded "waste" fractions contain proteins of potential interest and should be included in CSF biomarker studies.

Dataset Information

The challenge of increasing Pfam coverage of the human proteome.

Publications

The challenge of increasing Pfam coverage of the human proteome.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets