Mismatch string kernels for discriminative protein classification

By Christina S. Leslie, Eleazar Eskin, Adiel Cohen, Jason Aaron Edward Weston, W. S. Noble. 2004

592 citations recorded by the source corpus.

How researchers see this paper

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

In Molecular Biology

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].”
Biological Sequence Modeling with Convolutional Kernel Networks (2017, neutral): “Biological sequences are represented by a large set of descriptors, constructed for instance by Fisher score [Jaakkola et al., 2000], k-mer spectrum up to some mismatches [Leslie et al., 2003], or local alignment score [Saigo et al., 2004].”
Protein remote homology detection based on bidirectional long short-term memory (2017, methodology): “Several kernels have been proposed to calculate the similarity between protein samples, such as mismatch kernel [27], motif kernel [28], LA kernel [29], SW-PSSM [30], SVM-Pairwise [31], etc.”
Machine-learning-guided directed evolution for protein engineering (2018, background): “…kernels can be applied to protein inputs, there are also kernels designed for use on proteins, including spectrum and mismatch string kernels 30,31 , which count the number of shared subsequences between two proteins, and weighted decomposition kernels 32 , which account for three-dimensional…”
The language of proteins: NLP, machine learning & protein sequences (2021, background): “an enzyme, a receptor or a structural protein) [52,51], or where it is expressed in the cell (e.”
Profile-based string kernels for remote homology detection and motif extraction (2004, methodology): “Following the framework of k-mer based string kernels [16] [15] [17], our profile-based kernels will depend on a feature mapping to the |Σ| k -dimensional feature space indexed by the set of all possible k-length subsequences ( " k-mers " ) of amino acids, where k is a small positive integer.”
ROC analysis: applications to the classification of biological sequences and 3D structures (2008, methodology): “SVMs, artificial neural networks and other machine learning algorithms fall into this category.”
LocTree2 predicts localization for all domains of life (2012, methodology): “For the profile kernel, the matrix can be calculated very efficiently with the suffix tree-based ‘kernel trick’ introduced by the groups of Christina Leslie and Bill Noble (Leslie et al., 2004).”
Classification of real and pseudo microRNA precursors using local structure-sequence features and support vector machine (2005, methodology): “SVM has been widely applied to the prediction and classification of important biology signals such as promoters [16], translation initiation sites [17], splicing sites [18] and proteins [19].”
Protein-ligand interaction prediction: an improved chemogenomics approach (2008, background): “For example, the Dirac kernel only performs at an average AUC of 77% for the ion channel dataset, while the multitask kernel increases the AUC to 87.3% and the hierarchy kernel brings it to 92.5%.”
On Prediction Using Variable Order Markov Models (2004, methodology): “Previous studies on protein classiflcation used a variety of methods, including generative models such as HMMs (Sonnhammer et al., 1997), PSTs (Bejerano & Yona, 2001), and discriminative models such as Support Vector Machines (Ding & Dubchak, 2001; Leslie et al., 2004 ).”
Kernels for small molecules and the prediction of mutagenicity, toxicity and anti-cancer activity (2005, background): “Variations allow for word mismatches and insertions (Leslie et al., 2004).”
ProFET: Feature engineering captures high-level protein functions (2015, background): “The strongest features learned by the ML classifiers often expose biologically important motifs (Leslie et al. , 2004).”
DEEPre: sequence-based enzyme EC number prediction by deep learning (2017, background): “The third solution is to systematically generate sequence-length-independent features, such as string kernels (Dai et al. , 2017, Leslie et al. , 2002, 2004; Ra¨tsch et al. , 2005; Wang et al. , 2014), which, however, do not consider the classification problem when extracting features.”
Diet and dioxins: the need to cut back. (2004, methodology): “Kernel methods such as the support vector machine [10] have proven to give particularly good results, and several groups have introduced different types of kernel functions [11-19].”
Enhanced Regulatory Sequence Prediction Using Gapped k-mer Features (2014, methodology): “The mismatch string kernel [8] is one such method, originally motivated by the fact that homologous protein sequences are not usually identical and have many frequently mutated positions.”