SVM-Fold: a tool for discriminative multi-class protein fold and superfamily recognition.

SVM-Fold: a tool for discriminative multi-class protein fold and superfamily recognition. Research Abstract Details 

Research Abstract Table of Contents

Jump to the:

SVM-Fold: a tool for discriminative multi-class protein fold and superfamily recognition. Abstract Text:

Iain Melvin,Eugene Ie,Rui Kuang,Jason Weston,William Noble Stafford,Christina Leslie,

BACKGROUND: Predicting a protein's structural class from its amino acid sequence is a fundamental problem in computational biology. Much recent work has focused on developing new representations for protein sequences, called string kernels, for use with support vector machine (SVM) classifiers. However, while some of these approaches exhibit state-of-the-art performance at the binary protein classification problem, i.e. discriminating between a particular protein class and all other classes, few of these studies have addressed the real problem of multi-class superfamily or fold recognition. Moreover, there are only limited software tools and systems for SVM-based protein classification available to the bioinformatics community. RESULTS: We present a new multi-class SVM-based protein fold and superfamily recognition system and web server called SVM-Fold, which can be found at http://svm-fold.c2b2.columbia.edu. Our system uses an efficient implementation of a state-of-the-art string kernel for sequence profiles, called the profile kernel, where the underlying feature representation is a histogram of inexact matching k-mer frequencies. We also employ a novel machine learning approach to solve the difficult multi-class problem of classifying a sequence of amino acids into one of many known protein structural classes. Binary one-vs-the-rest SVM classifiers that are trained to recognize individual structural classes yield prediction scores that are not comparable, so that standard "one-vs-all" classification fails to perform well. Moreover, SVMs for classes at different levels of the protein structural hierarchy may make useful predictions, but one-vs-all does not try to combine these multiple predictions. To deal with these problems, our method learns relative weights between one-vs-the-rest classifiers and encodes information about the protein structural hierarchy for multi-class prediction. In large-scale benchmark results based on the SCOP database, our code weighting approach significantly improves on the standard one-vs-all method for both the superfamily and fold prediction in the remote homology setting and on the fold recognition problem. Moreover, our code weight learning algorithm strongly outperforms nearest-neighbor methods based on PSI-BLAST in terms of prediction accuracy on every structure classification problem we consider. CONCLUSION: By combining state-of-the-art SVM kernel methods with a novel multi-class algorithm, the SVM-Fold system delivers efficient and accurate protein fold and superfamily recognition.

SVM-Fold: a tool for discriminative multi-class protein fold and superfamily recognition. Publishing Authors By Initials

I Melvin,E Ie,R Kuang,J Weston,WN Stafford,C Leslie,

For similar information science: computing methodologies: software research abstracts see: information science: computing methodologies: software research

PUBMED ID PMID:

MEDLINE DATE:

SVM-Fold: a tool for discriminative multi-class protein fold and superfamily recognition. Journal Published:

PUBLICATION TYPE: Research Support, U.S. Gov't,

Journal: BMC bioinformatics

VOLUME: 8 Suppl 4

Page Numbers: S2

Journal Abbreviation: BMC Bioinformatics

ISSN: 1471-2105

DAY: 22

MONTH: 05

YEAR: 2007

SVM-Fold: a tool for discriminative multi-class protein fold and superfamily recognition. Information

Number of References:

LANGUAGE: eng

NlmUniqueID: 100965194

SVM-Fold: a tool for discriminative multi-class protein fold and superfamily recognition. Keywords Mesh Terms:

KEYWORDS: Software

MESH TERMS: methods

Chemical & Substance for Abstract: SVM-Fold: a tool for discriminative multi-class protein fold and superfamily recognition. Information

Substance Name: Proteins

Registry Number: 0

Grant and Affiliation Information for SVM-Fold: a tool for discriminative multi-class protein fold and superfamily recognition.

AFFILIATION: NEC Laboratories of America, Princeton, NJ, USA. imelvin@gmail.com

Country: England

AGENCY: United States NIGMS

GRANT: GM74257

ACRONYM: GM

MEDLINETA: BMC Bioinformatics

REFSOURCE:

DATABASENAME:

ACCESSION NUMBER:

Number Hits: 0

SVM-Fold: a tool for discriminative multi-class protein fold and superfamily recognition Related Publications

• 15-month-old infants detect violations in pretend scenarios.
• A common presentation with a rare cause.
• Age, sex and early environment contribute to individual differences in nicotine/acetaldehyde-induced behavioral and endocrine responses in rats.
• An evaluation of two indirect methods of estimating body weight in holstein calves and heifers.
• Anaesthesia for neurosurgery in the sitting position.
• Assessing penetration and aspiration: how do videofluoroscopy and fiberoptic endoscopic evaluation of swallowing compare?
• Challenges in improving fitness: results of a community-based, randomized, controlled lifestyle change intervention.
• Combining sequence and time series expression data to learn transcriptional modules.
• Consequences of the loss of p53, RB1, and PTEN: relationship to gefitinib resistance in endometrial cancer.
• Developmental regulation of nicotinic acetylcholine receptors within midbrain dopamine neurons.
• Gestational exposure to nicotine and monoamine oxidase inhibitors influences cocaine-induced locomotion in adolescent rats.
• Involvement of alpha1-adrenergic receptors in tranylcypromine enhancement of nicotine self-administration in rat.
• Living in a Glasshouse... embracing care issues beyond ICU.
• Low dose nicotine treatment during early adolescence increases subsequent cocaine reward.
• Models representing type I and type II human endometrial cancers: Ishikawa H and Hec50co cells.
• One world, one medicine.
• Ovarian function in Holstein cows immunized against pregnant mare serum gonadotrophin.
• Plasma concentrations of luteinizing hormone and progesterone during superovulation of dairy cows using follicle stimulating hormone or pregnant mare serum gonadotrophin.
• Predicting genetic regulatory response using classification.
• Relationship of estrogen and progesterone receptors to clinical outcome in metastatic endometrial carcinoma: a Gynecologic Oncology Group Study.
• Relative severity of aflatoxin contamination of cereal crops in West Africa.
• SVM-Fold: a tool for discriminative multi-class protein fold and superfamily recognition.
• Sample size requirements for bone density precision assessments and effect on patient categorization: a Monte Carlo simulation study.
• Stimulation of PI 3-kinase signaling via inhibition of the tumor suppressor phosphatase, PTEN.
• Synchronization of estrus and pregnancy risk in anestrous dairy cows after treatment with a progesterone-releasing intravaginal device.
• Targeted treatment using monoclonal antibodies and tyrosine-kinase inhibitors in pregnancy.
• The conceptual underpinnings of pretense: pretending is not 'behaving-as-if'.
• The role of advanced practice clinicians in the availability of abortion services in the United States.
• Timing of the onset and duration of ovulation in superovulated beef heifers.
• Type 1 diabetes and latent autoimmune diabetes in adults.