忐忑的铁板烧 · 石璐离开刺猬乐队,她的离开对乐队有何影响? ...· 1 年前 · |
销魂的香菜 · c# 实现网页(Web ...· 1 年前 · |
善良的拐杖 · 10. 動態組合 SQL 與執行語法 - ...· 1 年前 · |
气宇轩昂的红薯 · 基于 Spring Boot 2.5.1 ...· 1 年前 · |
害羞的鸵鸟
1 年前 |
Convolutional neural networks are a family of neural network architectures having at least one convolutional layer. LeNet is the original CNN network architecture bearing the name of Yann Lecun. Its architecture can be written as,
In this architecture, convolutional layer is the cornerstone of the CNN, which is a hidden layer where a square grid of weights is convolved with the input, just like an image mask. The output of the convolutional layer is akin to a convolved image. Next, the non-linear activation function, ReLu (REctified Linear Unit), is applied to zero-out any negative values. To reduce the dimension of the feature extracted from the convolutional layer, there is a pooling layer emulating downsampling . In general, each group of four values or pixels is replaced by the maximum (sometimes the mean) of the four, leaving a single most intense pixel. This pooling method is known as max pooling . This sequence of CONV->RELU->POOL layers may be repeated multiple times to create a deep architecture. Finally, a few fully-connected layers round off the architecture. Though it seems far more sophisticated than a MLP, it can be shown that a CNN can be represented as a classical fully-connected neural network. For example, a convolutional layer can be represented as a sparse fully-connected layer. Various techniques have been developed for training these vast models, for example momentum optimizers, weight initialization, batch normalization, and dropout.
Convolutional Neural Networks are the current state-of-the-art in many computer vision tasks. In addition to image classification, their great success has attracted wide attention in many fields. It has been found that using a pre-trained CNN as a general-purpose feature extractor for a simple linear model can yield significant improvements over even the most meticulously hand-crafted feature engineering.
The protein subcellular localization problem can be viewed as a multi-label multi-class classification task. Unlike the ordinary deep learning methods for multi-classification problems, in our method, we need to change the loss function. The most intuitive way is to extend the cross-entropy loss. The cross-entropy loss function is defined by
where Θ denotes the parameters of CNN model, is a set that contains the relevant localization of protein i and is the result for protein i on localization j , through a softmax activation:
Instead of using the cross-entropy loss function, the binary cross-entropy loss (BCE) over sigmoid activation has shown better performance when applied into multi-label task. The binary cross-entropy loss is
where
Tree boosting is a learning method to enhance the classification ability of weak classifiers by iteratively adding new decision trees to the ensembles of decision trees. Let denotes a dataset with n classes and m feature. Then the prediction of a tree boosting for a ( x i , y i ) is given by
where g j ( x i ) = w q ( x i ) is the prediction of the j -th decision tree with leaf weights w q on a datapoint x i , and M is the number of members in the ensemble.
It is well-known that the decision tree tends to overfit when the decision tree is fully grown. Thus, the set prediction function of decision trees g j can be learned by minimizing the objective function
where l i ( y i , ŷ i ) is a term which measures the goodness of the prediction ŷ i and the object y i . Ω( g j ) is a regularization term that does not depend on the data.
XGBoost implements a parallel tree boosting in a fast and accurate way. In XGBoost, the regularization function is chosen to be
with γ and λ regularization parameters that must be chosen appropriately. Notice this regularization penalizes both large weights on the leaves (similar to L 2 -regularization) and has large partitions.
As mentioned above, the tree boosting iteratively enlarges the ensemble of decision trees, then the prediction of the t -th iteration can be defined as
The objective function (7) at step t can be modified as
Apply a Taylor expansion on the objective function (10) to second order and then the final objective function at step t can be approximated as
where
Let j : I j = { i : q t ( x i ) = j } denotes the set of point x i mapped to leaf, and . Then we can rewrite the as
To find the optimal weight w j of leaf j for a fixed tree structure, q ( x ) can be obtained by applying the following equation
plugging back into gives
It is clear that measures the in-sample performance of g t and we should find the decision tree that minimizes this value. However, in practice, this is impossible to enumerate all possible trees over the data and find the tree which can minimize . Instead, an approximate greedy algorithm runs to optimize one level of the tree at a time by trying to find optimal splits of the data, leading to a tree with a local minimum of , which is then added to the ensemble.
For the multi-label multi-class classification problem, we utilize XGBoost as classifiers and adopt the binary relevance strategy (Boutell et al., 2004 ) to construct m binary classifiers.
Figure Figure1 1 gives the overall structure of the CNN-XGBoost model for protein subcellular location prediction. The input of the model is a one-dimensional vector and constructed by the position specific scoring matrices (PSSM) and proteins interaction scoring matrix which are extracted from STRING and GO terms semantic similarities. On this basis, a protein can be expressed as L × 1 vector ( L is the number of sequences in training set), analog image data equivalent to a protein is a one-dimensional “image” with 1 channels. So the input is a L × 1 matrix.
After obtaining the proper feature representations by the trained CNN, compared with the classic CNN, our CNN-XGBoost model replaces the soft-max layer of CNN with XGBoost to predict the localization of subcellular of proteins, which enables features automatically obtained from input and provides more precise and efficient classification.
To verify the performance of our method, we employ three protein datasets: the Hum-mPloc3.0, the BaCelLo animals, and the Hoglund. Table Table1 1 gives the details of these datasets. The train set of Hum-mPloc 3.0 consists of 3,122 proteins and 1,023 proteins own more than one label. The test set of Hum-mPloc 3.0 consists of 379 proteins, among which 120 proteins belong to multi-label proteins. Each protein in Hum-mPloc 3.0 is assigned at least one label of 12 subcellular locations (Centrosome, Cytoplasm, Cytoskeleton, Endoplasmic reticulum, Endosome, Extracellular, Golgi apparatus, Lysosome, Mitochondrion, Nucleus, Peroxisome, and Plasma membrane).
Hum-mLoc 3.0 | BaCelLo | Hoglund | ||||
---|---|---|---|---|---|---|
Training | Testing | Training | Testing | Training | Testing | |
No. Proteins | 3,126 | 379 | 2,597 | 576 | 5,959 | 158 |
No. Labels | 4,229 | 541 | 2,597 | 576 | 5,959 | 158 |
No.Locations | 12 | 4 | 6 |
For the BaCelLo dataset, there are four subcellular locations: Cytoplasm, Mitochondrion, Nucleus, and Secreted. The size of the training set is set to 2,597 and the testing set consists of 576 proteins. All the proteins of BaCelLo dataset are of a single label. In the Hoglund dataset, the training set includes nine subcellular locations (Nucleus, Cytoplasm, Mitochondrion, Endoplasmic reticulum, Golgi apparatus, Peroxisome, Plasma membrane, Extracellular space, Lysosome, and Vacuole), and the test consists of 158 proteins with six subcellular locations (Endoplasmic reticulum, Golgi apparatus, Peroxisome, Plasma membrane, Extracellular space, and Lysosome).
A widely-applied method for evaluating a mutli-label multi-class classifier is to compute the ACC and F1 values. ACC is the average of ACC x i of all proteins in the testing set, calculated for protein x i is
where TP, FP, and FN are true positive, false positive, and false negative, respectively. The F1 score considers both the harmonic mean of precision and recall of subcellular location y j , defined as follows:
where T j and P j are the set of proteins for true location y j and the set of proteins for predicted locations y j respectively.
To verify the performance of our approach, some typical protein subcellular location tools including Hum-mPLoc 3.0 (Zhou et al., 2016 ), YLoc+ (Briesemeister et al., 2010 ), iLoc-Hum (Chou et al., 2012 ), WegoLoc (Chi and Nam, 2012 ), mLASSO-Hum (Wan et al., 2015 ), and PSL-Recommender (Jamali et al., 2018 ) were compared to our method. The F1 score and ACC for each subcellular localization are summarized in Table Table2 2 and Figure Figure2 2 for Hum-mploc 3.0 dataset. As seen in Table Table2 2 and Figure Figure2, 2 , the CNN-XGBoost outperforms the mean value of F1 score and ACC of all other methods. Also, in 7 out of 12 subcellular locations, CNN-XGBoost has the best performance among all the methods while in the other three locations it has the second best performance. It is only in centrosome and endosome that CNN-XGBoost shows unsatisfactory results. As seen in Table Table3, 3 , the CNN-XGBoost slightly outperforms the second best method by both mean F1 score and ACC.
Location | iLoc-Human | WegoLoc | mLASSO-Hum | Hum-mLoc 3.0 | PSL-Recommender | CNN-XGBoost | ||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
pre | re | F1 | pre | re | F1 | pre | re | F1 | pre | re | F1 | pre | re | F1 | pre | re | F1 | |
Centrosome | 0 | 0 | 0 | 0.75 | 0.14 | 0.23 | 0.59 | 0.59 | 0.59 | 0.75 | 0.55 | 0.63 | 0.94 | 0.75 | 0.83 | 0.79 | 0.50 | 0.61 |
Cytoplasm | 0.5 | 0.54 | 0.52 | 0.69 | 0.53 | 0.60 | 0.93 | 0.51 | 0.66 | 0.76 | 0.73 | 0.74 | 0.79 | 0.81 | 0.80 | 0.85 | 0.89 | 0.87 |
Cytoskeleton | 0 | 0 | 0 | 0.32 | 0.34 | 0.33 | 0.9 | 0.22 | 0.35 | 0.8 | 0.68 | 0.74 | 0.93 | 0.77 | 0.84 | 0.89 | 0.80 | 0.85 |
ER | 0 | 0 | 0 | 0.73 | 0.2 | 0.31 | 0.74 | 0.49 | 0.59 | 0.83 | 0.37 | 0.51 | 0.9 | 0.7 | 0.79 | 0.97 | 0.71 | 0.82 |
Endosome | 0 | 0 | 0 | 0.25 | 0.07 | 0.11 | 0.38 | 0.2 | 0.26 | 0.58 | 0.47 | 0.52 | 0.57 | 0.37 | 0.45 | 0.80 | 0.27 | 0.40 |
Extracellular | 0.62 | 0.62 | 0.62 | 0.67 | 0.77 | 0.71 | 0.16 | 0.69 | 0.26 | 0.5 | 0.46 | 0.48 | 0.66 | 0.71 | 0.68 | 0.80 | 0.62 | 0.70 |
Golgi apparatus | 0.6 | 0.3 | 0.4 | 0.6 | 0.15 | 0.24 | 0.72 | 0.65 | 0.68 | 0.69 | 0.45 | 0.55 | 0.88 | 0.61 | 0.72 | 0.80 | 0.60 | 0.69 |
Lysosome | 0.5 | 0.13 | 0.2 | 0.2 | 0.13 | 0.15 | 0.55 | 0.75 | 0.63 | 0.71 | 0.63 | 0.67 | 1 | 0.55 | 0.71 | 1.00 | 0.75 | 0.86 |
Mitochondrion | 0.95 | 0.33 | 0.49 | 0.79 | 0.73 | 0.76 | 0.83 | 0.88 | 0.85 | 0.78 | 0.75 | 0.76 | 0.92 | 0.88 | 0.90 | 0.96 | 0.80 | 0.87 |
Nucleus | 0.54 | 0.7 | 0.61 | 0.65 | 0.64 | 0.64 | 0.85 | 0.7 | 0.76 | 0.75 | 0.71 | 0.73 | 0.81 | 0.92 | 0.87 | 0.83 | 0.91 | 0.87 |
Peroxisome | 1 | 0.5 | 0.67 | 0.5 | 1 | 0.67 | 0.29 | 1 | 0.44 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 |
Plasma membrane | 0.42 | 0.33 | 0.37 | 0.44 | 0.53 | 0.48 | 0.58 | 0.56 | 0.57 | 0.65 | 0.44 | 0.52 | 0.78 | 0.74 | 0.76 | 0.89 | 0.75 | 0.81 |
ACC-mean | 0.41 | 0.50 | 0.65 | 0.63 | 0.77 | 0.78 | ||||||||||||
F1-mean | 0.32 | 0.44 | 0.56 | 0.65 | 0.78 | 0.80 |
The bold marks the first best result and the underline marks the second best result .
BaCelLo | Hoglund | |
---|---|---|
MultiLoc2-LowRes | 0.73/0.76 | – |
MultiLoc2-HighRes | 0.68/0.71 | 0.57/0.41 |
BaCelLo | 0.64/0.66 | – |
Hum-mPloc 3.0 | 0.86/0.84 | 0.64/0.59 |
PSL-Recommender | 0.94/0.92 | 0.92/0.90 |
CNN-XGBoost | 0.94/0.94 | 0.94/0.92 |
The bold marks the first best result and the underline marks the second best result .
In addition, we also evaluated our method on the DeepLoc dataset, compared to the DeepLoc, our method provides slightly better prediction with significantly lighter model, meanwhile, it is known that DeepLoc can not handle multilabel multiclass problem, whereas our method still shows outstanding performance.
In order to make balance of the classification performance and the complexity when training the model for the protein subcellular location in Alzheimer's disease, this paper proposes a prediction framework integrating CNN and XGBoost, taking advantage of the outstanding ability of feature expression of CNN, and the good classification performance of XGBoost. Experiments are implemented on the Hum-mPloc3.0, the BaCelLo animals, and the Hoglund database, and the results demonstrate that the new method outperforms the typical machine learning based tools. Further work will focus on the verification of our model on more datasets, especially the datasets related to Alzheimer's disease, and the optimization of the structure of CNN utilized in the model.
All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
This work is supported by the National Natural Science Foundation of China (NSFC, Grant no. 61305013 and 61872114). The authors would like to thank the reviewers for their valuable comments.
忐忑的铁板烧 · 石璐离开刺猬乐队,她的离开对乐队有何影响? - 知乎 1 年前 |