Data Augmentation Algorithms for Detecting Conserved Domains in Protein Sequences: A Comparative Study

Improved Chemical Structure-Activity Modeling Through Data Augmentation

Extending the original training data with simulated unobserved data points has proven powerful to increase both the generalization ability of predictive models and their robustness against changes in the structure of data (e.g., systematic drifts in the response variable) in diverse areas such as the analysis of spectroscopic data or the detection of conserved domains in protein sequences. In this contribution, we explore the effect of data augmentation in the predictive power of QSAR models, quantified by the RMSE values on the test set. We collected 8 diverse data sets from the literature and ChEMBL version 19 reporting compound activity as pIC50 values. The original training data were replicated (i.e., augmented) N times (N ∈ 0, 1, 2, 4, 6, 8, 10), and these replications were perturbed with Gaussian noise (μ = 0, σ = σnoise) on either (i) the pIC50 values, (ii) the compound descriptors, (iii) both the compound descriptors and the pIC50 values, or (iv) none of them. The effect of data augmentation was evaluated across three different algorithms (RF, GBM, and SVM radial) and two descriptor types (Morgan fingerprints and physicochemical-property-based descriptors). The influence of all factor levels was analyzed with a balanced fixed-effect full-factorial experiment. Overall, data augmentation constantly led to increased predictive power on the test set by 10-15%. Injecting noise on (i) compound descriptors or on (ii) both compound descriptors and pIC50 values led to the highest drop of RMSEtest values (from 0.67-0.72 to 0.60-0.63 pIC50 units). The maximum increase in predictive power provided by data augmentation is reached when the training data is replicated one time. Therefore, extending the original training data with one perturbed repetition thereof represents a reasonable trade-off between the increased performance of the models and the computational cost of data augmentation, namely increase of (i) model complexity due to the need for optimizing σnoise and (ii) the number of training examples.

A Monte Carlo EM algorithm for de novo motif discovery in biomolecular sequences

Motif discovery methods play pivotal roles in deciphering the genetic regulatory codes (i.e., motifs) in genomes as well as in locating conserved domains in protein sequences. The Expectation Maximization (EM) algorithm is one of the most popular methods used in de novo motif discovery. Based on the position weight matrix (PWM) updating technique, this paper presents a Monte Carlo version of the EM motif-finding algorithm that carries out stochastic sampling in local alignment space to overcome the conventional EM's main drawback of being trapped in a local optimum. The newly implemented algorithm is named as Monte Carlo EM Motif Discovery Algorithm (MCEMDA). MCEMDA starts from an initial model, and then it iteratively performs Monte Carlo simulation and parameter update until convergence. A log-likelihood profiling technique together with the top-k strategy is introduced to cope with the phase shifts and multiple modal issues in motif discovery problem. A novel grouping motif alignment (GMA) algorithm is designed to select motifs by clustering a population of candidate local alignments and successfully applied to subtle motif discovery. MCEMDA compares favorably to other popular PWM-based and word enumerative motif algorithms tested using simulated (l, d)-motif cases, documented prokaryotic, and eukaryotic DNA motif sequences. Finally, MCEMDA is applied to detect large blocks of conserved domains using protein benchmarks and exhibits its excellent capacity while compared with other multiple sequence alignment methods.