Image-based promoter prediction: a promoter prediction method based on evolutionarily generated patterns

Abstract

Prediction of promoter regions is crucial for studying gene function and regulation. The well-accepted position weight matrix method for this purpose relies on predefined motifs, which would hinder application across different species. Here, we introduce image-based promoter prediction (IBPP) as a method that creates an “image” from training promoter sequences using an evolutionary approach and predicts promoters by matching with the “image”. We used Escherichia coli σ70 promoter sequences to test the performance of IBPP and the combination of IBPP and a support vector machine algorithm (IBPP-SVM). The “images” generated with IBPP could effectively distinguish promoter from non-promoter sequences. Compared with IBPP, IBPP-SVM showed a substantial improvement in sensitivity. Furthermore, both methods showed good performance for sequences of up to 2,000 nt in length. The performances of IBPP and IBPP-SVM were largely affected by the threshold and dimension of vectors, respectively. The source code and documentation are freely available at https://github.com/hahatcdg/IBPP.

Novel transformer networks for improved sequence labeling in genomics

“…Support vector machines are applied by Manavalan et al to predict phage virion proteins present in the bacterial genome [25]. Further examples of the application of support vector machines include the work of: Goel et al [12], who propose an improved method for splice site prediction in Eukaryotes; and, Wang et al [36], who introduce the detection of σ 70 promoters using evolutionary driven image creation.…”

“…Differences exist between the negative sets, as our method processes the full genome. This is in contrast to the use of a subsampled negative set by recent studies [29,24,36,23,34,6,29]. In the majority of studies, custom sampling methods are used, where the samples of the negative set is not publicly available [23,29,34,17,6].…”

Evaluating the Potential of Applying Machine Learning Tools to Metabolic Pathway Optimization

“…But, if promoter sequences in a training dataset are accurate, this would reduce the noise in the model and afford more accurate prediction. Hence, the computational challenge in applying machine learning to promoter strength prediction lies in the identification of small snippets of nucleotide sequence that strongly correlates with expression level [33]. Currently, a commonly used method for extracting sequence features is position weight matrix [34], but the approach may not be transferable to different species [33].…”

Explainable Transformer Models for Functional Genomics in Prokaryotes

“…Given its importance to the community, TSSs annotated with at least 'strong' evidence are retrieved from the RegulonDB database, summing to 6,487 annotated positions. RegulonDB features an upto-date collection of manual and automated TSSs collected from a plethora of independent sources [27], and has been the chosen data set for recent machine learning implementations [23,19,35,18].…”

“…As the nucleotide sequence downstream of the TSS is expected to be of relevance to its identification, labels can be shifted downstream in order to make the information of these regions accessible to the model. In this study, all labels have been shifted downstream by 20 nucleotides, a distance that is used by previous computational methods [23,19,35,18]. Studies on the mechanisms of RNAP binding do furthermore not describe interactions of the RNAP binding process beyond position +20 of the TSS [13,26].…”