This is an implementation of ULMFiT. The model architecture used is based on the AWD-LSTM.
The approach uses three training phases to produce a classification model:
- Train a language model on a large, unlabeled corpus
- Fine tune the language model on the classification corpus
- Use the fine tuned language model to initialize a classification model
This is useful for genetic analysis, where large amounts of unlabeled data is abundant and labeled data is scarce. The approach allows to train a model on a large, unlabeled genomic corpus in an unsupervised fashion. The pre-trained language model serves as a feature extractor for parsing genomic data.
The method performs well at the task of classifying promoter sequences from random sections of the genome. The process of unsupervised pre-training and fine-tuning has a clear impact on the performance of the classification model
Model | Accuracy | Precision | Recall | Correlation Coefficient |
---|---|---|---|---|
Naive | 0.834 | 0.847 | 0.816 | 0.670 |
E. coli Genome Pre-Training | 0.919 | 0.941 | 0.893 | 0.839 |
Genomic Ensemble Pre-Training | 0.973 | 0.980 | 0.966 | 0.947 |
Classification performance on human promoters is competitive with published results
For the short promoter sequences, using data from Recognition of Prokaryotic and Eukaryotic Promoters using Convolutional Deep Learning Neural Networks:
Model | DNA Size | kmer/stride | Accuracy | Precision | Recall | Correlation Coefficient | Specificity |
---|---|---|---|---|---|---|---|
Kh et al. | -200/50 | - | - | - | 0.9 | 0.89 | 0.98 |
Naive Model | -200/50 | 5/2 | 0.80 | 0.74 | 0.80 | 0.59 | 0.80 |
With Pre-Training | -200/50 | 5/2 | 0.922 | 0.963 | 0.849 | 0.844 | 0.976 |
With Pre-Training and Fine Tuning | -200/50 | 5/2 | .977 | .959 | .989 | .955 | .969 |
With Pre-Training and Fine Tuning | -200/50 | 5/1 | .990 | .983 | .995 | .981 | .987 |
With Pre-Training and Fine Tuning | -200/50 | 3/1 | .995 | .992 | .996 | .991 | .994 |
For the long promoter sequences, using data from PromID: Human Promoter Prediction by Deep Learning:
Model | DNA Size | Models | Accuracy | Precision | Recall | Correlation Coefficient |
---|---|---|---|---|---|---|
Umarov et al. | -1000/500 | 2 Model Ensemble | - | 0.636 | 0.802 | 0.714 |
Umarov et al. | -200/400 | 2 Model Ensemble | - | 0.769 | 0.755 | 0.762 |
Naive Model | -500/500 | Single Model | 0.858 | 0.877 | 0.772 | 0.708 |
With Pre-Training | -500/500 | Single Model | 0.888 | 0.90 | 0.824 | 0.770 |
With Pre-Training and Fine Tuning | -500/500 | Single Model | 0.892 | 0.877 | 0.865 | 0.778 |
This table shows results on data from Recognition of prokaryotic and eukaryotic promoters using convolutional deep learning neural networks.
Method | Organism | Training Examples | Accuracy | Precision | Recall | Correlation Coefficient | Specificity |
---|---|---|---|---|---|---|---|
Kh et al. | E. coli | 2936 | - | - | 0.90 | 0.84 | 0.96 |
ULMFiT | E. coli | 2936 | 0.956 | 0.917 | 0.880 | 0.871 | 0.977 |
Kh et al. | B. subtilis | 1050 | - | - | 0.91 | 0.86 | 0.95 |
ULMFiT | B. subtilis | 1050 | 0.905 | 0.857 | 0.789 | 0.759 | 0.95 |
ULMFiT shows improved performance on the metagenomics taxonomic dataset from Deep learning models for bacteria taxonomic classification of metagenomic data.
Method | Data Source | Accuracy | Precision | Recall | F1 |
---|---|---|---|---|---|
Fiannaca et al. | Amplicon | .9137 | .9162 | .9137 | .9126 |
ULMFiT | Amplicon | .9239 | .9402 | .9332 | .9306 |
Fiannaca et al. | Shotgun | .8550 | .8570 | .8520 | .8511 |
ULMFiT | Shotgun | .8797 | .8824 | .8769 | .8758 |
Trained on a dataset of mammalian enhancer sequences from Enhancer Identification using Transfer and Adversarial Deep Learning of DNA Sequences, ULMFiT outperforms Cohn et al.
Model - AUROC | Human | Mouse | Dog | Opossum |
---|---|---|---|---|
Cohn et al. | 0.80 | 0.78 | 0.77 | 0.72 |
ULMFiT 5-mer Stride 2 | 0.812 | 0.871 | 0.773 | 0.787 |
ULMFiT 4-mer Stride 2 | 0.804 | 0.876 | 0.771 | 0.786 |
ULMFiT 3-mer Stride 1 | 0.819 | 0.875 | 0.788 | 0.798 |
This table shows results for training a classification model on a dataset of coding mRNA sequences and long noncoding RNA (lncRNA) sequences. The dataset used is A deep recurrent neural network discovers complex biological rules to decipher RNA protein-coding potential by Hill et al.
Model | Test Set | Accuracy | Specificity | Sensitivity | Precision | MCC |
---|---|---|---|---|---|---|
GRU Ensemble (Hill et al.) | Standard Test Set | 0.96 | 0.97 | 0.95 | 0.97 | 0.92 |
ULMFiT (3mer stride 1) | Standard Test Set | 0.963 | 0.952 | 0.974 | 0.953 | 0.926 |
GRU Ensemble (Hill et al.) | Challenge Test Set | 0.875 | 0.95 | 0.80 | 0.95 | 0.75 |
ULMFiT (3mer stride 1) | Challenge Test Set | 0.90 | 0.944 | 0.871 | 0.939 | 0.817 |
In the plot below, the red line corresponds to a true transcription start site. The plot shows how prediction results are sensitive to changes around that location. Model Interpretations directory.
The image below shows a sample prediction of promoter locations on a 40,000 bp region of the E. coli genome. True promoter locations are shown in red.notebook