Enhanced Regulatory Sequence Prediction Using Gapped k-mer Features

By Mahmoud Ghandi, Dongwon Lee, Morteza Mohammad Noori, Michael A. Beer. 2014

497 citations recorded by the source corpus.

Abstract Oligomers of length k, or k-mers, are convenient and widely used features for modeling the properties and functions of DNA and protein sequences. However, k-mers suffer from the inherent limitation that if the parameter k is increased to resolve longer features, the probability of observing any specific k-mer becomes very small, and k-mer counts approach a binary variable, with most k-mers absent and a few present once. Thus, any statistical learning approach using k-mers as features becomes susceptible to noisy training set k-mer frequencies once k becomes large. To address this problem, we introduce alternative feature sets using gapped k-mers, a new classifier, gkm-SVM, and a general method for robust estimation of k-mer frequencies. To make the method applicable to large-scale genome wide applications, we develop an efficient tree data structure for computing the kernel matrix. We show that compared to our original kmer-SVM and alternative approaches, our gkm-SVM predicts functional genomic regulatory elements and tissue specific enhancers with significantly improved accuracy, increasing the precision by up to a factor of two. We then show that gkm-SVM consistently outperforms kmer-SVM on human ENCODE ChIP-seq datasets, and further demonstrate the general utility of our method using a Naïve-Bayes classifier. Although developed for regulatory sequence analysis, these methods can be applied to any sequence classification problem.

Subfields: Molecular Biology

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

Big Bird: Transformers for Longer Sequences (2020, neutral): “Table 7: Chromatin-Profile Prediction Model TF HM DHS gkm-SVM [30] 89.”
DanQ: a hybrid convolutional and recurrent deep neural network for quantifying the function of DNA sequences (2015, background): “It significantly outperforms gkmSVM (8), a related algorithm that can also predicts the regulatory function of DNA sequences, but uses a support vector machine instead of a CNN for predictions.”
iEnhancer-2L: a two-layer predictor for identifying enhancers and their strength by pseudo k-tuple nucleotide composition (2016, methodology): “Although these rates are only slightly better than the corresponding jackknife rates obtained by gkm-SVM (Ghandi et al., 2014) and kmer (Ioshikhes et al., 1996) on the same benchmark dataset used in this study, the success rates (see the third row of Table 1) achieved by the second layer of iEnhance”
Deep learning improves antimicrobial peptide recognition (2018, methodology): “To ensure that findings are applicable beyond DNNs, we repeat the DNN-reduced versus randomized alphabet experiment described above using gkmSVM vr.2.0 (Ghandi et al. , 2014).”
SBSM-Pro: support bio-sequence machine for proteins (2023, unknown): “Notably, gkmSVM was one of the ﬁrst to use kernel methods for biological sequence predictions, with the most common frequency feature being k-mer, and yielded promising results in certain scenarios, such as predicting enhancer activity in speciﬁc cell types [19] and disease-relevant mutations [20].”
A method to predict the impact of regulatory variants from DNA sequence (2015, background): “When trained on a set of putative regulatory sequences, our established gapped k -mer support vector machine (gkm-SVM) 7 identifies sequence features within these regulatory regions that determine their cell type–dependent activity.”
In Silico Predictive Modeling of CRISPR/Cas9 guide efficiency (2015, background): “Additional features such as those based on string mismatch kernels 33 , gapped kmers 34 and structural information of the guide/gene 18 could also prove useful.”
Deep Motif Dashboard: Visualizing and Understanding Genomic Sequences Using Deep Neural Networks (2016, methodology): “Later, SVM techniques outperformed the generative methods by using k-mer features [6, 20], but string kernel based SVM systems are limited by expensive computational cost proportional to the number of training and testing sequences.”
DeepEnhancer: Predicting enhancers by convolutional neural networks (2016, methodology): “improved kmer-SVM by introducing an alternative feature sets named gapped k-mers [11].”
Convolutional Kitchen Sinks for Transcription Factor Binding Site Prediction (2017, methodology): “We note that decreasing γ in Equation 1 relaxes the penalty of n -gram mismatches between disappoints, thereby capturing the behavior of the mismatch and gapped n -gram kernels [3, 13].”
Gkmexplain: Fast and Accurate Interpretation of Nonlinear Gapped k-mer SVMs Using Integrated Gradients (2018, methodology): “Because the feature space corresponding to gapped k -mer counts can be quite large, a more computationally efﬁcient formulation [11, 6] of the gapped k -mer kernel performs a sum only over the full l -mers that are actually present in the sequences.”
SeqGL Identifies Context-Dependent Binding Signals in Genome-Wide Regulatory Element Maps (2015, methodology): “Several other k-mer based discriminative models have recently been proposed to learn TF binding preferences from ChIP-seq data, including two SVMmethods: the di-mismatch kernel, based on k-mer features in the dinucleotide alphabet counted with mismatches [14]; and the gkm-SVMmethod [20], which is ve”
Basset: learning the regulatory code of the accessible genome with deep convolutional neural networks (2015, background): “3 from http://www.beerlab.org/gkmsvm/ (Lee et al. 2015; Ghandi et al. 2014).”
Sequential regulatory activity prediction across chromosomes with convolutional neural networks (2017, background): “Successful predictive modeling of transcription factor (TF) binding, accessible chromatin, and histone modifications has provided mechanistic insight and useful interpretation of genomic variants (Whitaker et al. 2015; Ghandi et al. 2014; Alipanahi et al. 2015; Zhou and Troyanskaya 2015; Kelley et a”
Machine Learning for Integrating Data in Biology and Medicine: Principles, Practice, and Opportunities (2018, background): “For example, gkm-SVM [132, 133] ﬁnds short sequences (k-mers) that are indicative of enhancer activity.”