Using the Fisher Kernel Method to Detect Remote Protein Homologies

By Tommi S. Jaakkola, Mark Diekhans, David Haussler. 1999

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

MRFalign: Protein Homology Detection through Alignment of Markov Random Fields (2014, methodology): “Improvement in these alignment-free methods results from the application of discriminative learning approaches such as SVM– Fisher [26], SVM-pairwise [27], SVM with the spectrum kernel [28] and SVM with the mismatch kernel [29].”
Task2Vec: Task Embedding for Meta-Learning (2019, background): “Variants of the Fisher kernel have found wide use as a representation of images [28, 29], and other structured data such as protein molecules [17] and text [30].”
Kernels for small molecules and the prediction of mutagenicity, toxicity and anti-cancer activity (2005, methodology): “Thus there is room for developing, testing and combining new 1D, 2D and 3D kernels and other kernels, such as Fisher kernels (Jaakkola et al., 1999).”
Online handwriting recognition with support vector machines - a kernel approach (2002, methodology): “[10] developed an SVM kernel in their application of protein homology detection and refer to it as Fisher kernel.”
ROC analysis: applications to the classification of biological sequences and 3D structures (2008, methodology): “Generative models (consensus representations) Sequence group Nearest neighbor Sequence alilgnment score [24, 60] Multiple alignment HMM [40]”
Detecting Remote Evolutionary Relationships among Proteins by Large-Scale Semantic Embedding (2011, background): “In evaluations of remote homology detection algorithms, some researchers prefer to ignore members of the same family as the query, since these family members are presumably easy to identify [18].”
Convolutional Kitchen Sinks for Transcription Factor Binding Site Prediction (2017, methodology): “String kernel methods are well understood and have been extensively used for sequence classification [2, 3, 4].”
Neural Network-based Graph Embedding for Cross-Platform Binary Code Similarity Detection (2017, methodology): “Two representative examples are the Fisher kernel [23] and the probability product kernel [24].”
Fast model-based protein homology detection without alignment (2007, methodology): “The best-performing methods would take about 25 days on present-day computers to classify the sequences of a new genome (20 000 genes) as belonging to just one specific class—however, there are hundreds of classes.”
Support vector machine classification and validation of cancer tissue samples using microarray expression data (2000, methodology): “…of the SVM classifier which can be obtained at http://www.cs.columbia.edu/ ∼bgrundy/svm. † This implementation is based on that described in Jaakkola et al. (1999) and differs slightly from the above explanation in that it does not include a bias term, b, forcing all decision boundaries…”
Adaptive computation and machine learning (1998, methodology): “In [275] a general technique is presented for combining kernel methods with probabilistic generative models.”
Multi-class protein fold recognition using support vector machines and neural networks (2001, methodology): “, 1998); they are now extended to problems involving more than two classes, through the one-versus-others method (e.g. Dubchak et al., 1999; Jaakkola et al., 1999).”
Recognition of Structure Classification of Protein Folding by NN and SVM Hierarchical Learning Architecture (2003, methodology): “The SVM method, which has the advantage of fast convergence, was combined with the decision tree algorithm for multi-class protein folds recognition in order to get higher prediction accuracy [8].”
Probabilistic multi-class multi-kernel learning: on protein fold recognition and remote homology detection (2008, background): “The original dataset from Ding and Dubchak (2001) (based on SCOP PDB-40D) consists of 313 proteins for training and 385 proteins for testing with 5 35% sequence identity between any two proteins in the train and the test set.”
A Meta-Learning Method to Select the Kernel Width in Support Vector Regression (2004, methodology): “A heuristic to set the width of the Gaussian kernel for SVM has been used by Jaakkola, Diekhans, and Haussler (1999) .”
Gene functional classification from heterogeneous data (2001, methodology): “We also plan to experiment with the Fisher kernel method, in which each type of data is compared to a probab ilistic model of the domain [13, 12].”