There is, I think, something ethereal about i —the square root of minus one. I remember first hearing about it at school. It seemed an odd beast at that time—an intruder hovering on the edge of reality.

Usually familiarity dulls this sense of the bizarre, but in the case of i it was the reverse: over the years the sense of its surreal nature intensified. It seemed that it was impossible to write mathematics that described the real world in …

I and i

Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

[신간의 별자리x] 우리/미술, 그리고 ‘슬픔의 박물관’

Data Mining Practical Machine Learning Tools and Techniques

We benchmarked SignalP 4.0 against SignalP 3.0 and ten other signal peptide prediction algorithms (Fig. 1). We compared prediction performance using the Matthews correlation coefficient16, for which each sequence was counted as a true or false positive or negative. To test SignalP 4.0 performance, we did not use data that had been used in training the networks or selecting the optimal architecture, and the test data did not contain homologs to the training and optimization data (Supplementary Methods). The test set for SignalP 3.0 was also independent of the training set because we removed sequences used to construct SignalP 3.0 and their homologs from the benchmark data. For other algorithms more recent than SignalP 3.0, the benchmark data may include data used to train the methods, possibly leading to slight overestimations of their performance. Our results show that SignalP 4.0 was the best signal-peptide predictor for all three organism types (Fig. 1). This comes at a price, however, because SignalP 4.0 was not in all cases as good as SignalP 3.0 according to cleavage-site sensitivity or signal-peptide correlation when there are no transmembrane proteins present (Supplementary Results). An ideal method would have the best SignalP 4.0: discriminating signal peptides from transmembrane regions

SignalP 4.0: discriminating signal peptides from transmembrane regions

The cellular attributes of a protein, such as which compartment of a cell it belongs to and how it is associated with the lipid bilayer of an organelle, are closely correlated with its biological functions. The success of human genome project and the rapid increase in the number of protein sequences entering into data bank have stimulated a challenging frontier: How to develop a fast and accurate method to predict the cellular attributes of a protein based on its amino acid sequence? The existing algorithms for predicting these attributes were all based on the amino acid composition in which no sequence order effect was taken into account. To improve the prediction quality, it is necessary to incorporate such an effect. However, the number of possible patterns for protein sequences is extremely large, which has posed a formidable difficulty for realizing this goal. To deal with such a difficulty, the pseudo-amino acid composition is introduced. It is a combination of a set of discrete sequence correlation factors and the 20 components of the conventional amino acid composition. A remarkable improvement in prediction quality has been observed by using the pseudo-amino acid composition. The success rates of prediction thus obtained are so far the highest for the same classification schemes and same data sets. It has not escaped from our notice that the concept of pseudo-amino acid composition as well as its mathematical framework and biochemical implication may also have a notable impact on improving the prediction quality of other protein features.

https://onlinelibrary.wiley.com/doi/pdf/10.1002/prot.1072

Prediction of protein cellular attributes using pseudo‐amino acid composition

Some remarks on protein attribute prediction and pseudo amino acid composition.

A protein is usually classified into one of the following five struc- tural classes: a!, j3, a! +j3, a!/j3, and ( (irregular). The structural class of aprotein is correlated with its amino acid composition. However, given the amino acid composition of aprotein, how may one predict its structural class? Various efforts have been made in addressing this problem. This review addresses the progress in this field, with the focus on the state of the art, which is featured by a novel prediction algorithm and a recently developed database. The novel algorithm is characterized by a covariance matrix that takes into account the coupling effect among different amino acid components of a protein. The new database was established based on the requirement that the classes should have (1) as many nonhomologous structures as possible, (2) good quality structure, and (3) typical or distinguishable features for each of the structural classes concerned. The very high success rate for both the training-set proteins and the testing-set proteins, which has been further validated by a simulated analysis and a jackknife analysis, indicates that it is possible to predict the structural class of a protein according to its amino acid composition if an ideal and complete database can be established. It also suggests that the overall fold of a protein is basically determined by its amino acid composition.

Prediction of protein structural classes.

Information on subcellular localization of proteins is important to molecular cell biology, proteomics, system biology and drug discovery. To provide the vast majority of experimental scientists with a user-friendly tool in these areas, we present a package of Web servers developed recently by hybridizing the 'higher level' approach with the ab initio approach. The package is called Cell-PLoc and contains the following six predictors: Euk-mPLoc, Hum-mPLoc, Plant-PLoc, Gpos-PLoc, Gneg-PLoc and Virus-PLoc, specialized for eukaryotic, human, plant, Gram-positive bacterial, Gram-negative bacterial and viral proteins, respectively. Using these Web servers, one can easily get the desired prediction results with a high expected accuracy, as demonstrated by a series of cross-validation tests on the benchmark data sets that covered up to 22 subcellular location sites and in which none of the proteins included had > or =25% sequence identity to any other protein in the same subcellular-location subset. Some of these Web servers can be particularly used to deal with multiplex proteins as well, which may simultaneously exist at, or move between, two or more different subcellular locations. Proteins with multiple locations or dynamic features of this kind are particularly interesting, because they may have some special biological functions intriguing to investigators in both basic research and drug discovery. This protocol is a step-by-step guide on how to use the Web-server predictors in the Cell-PLoc package. The computational time for each prediction is less than 5 s in most cases. The Cell-PLoc package is freely accessible at http://chou.med.harvard.edu/bioinf/Cell-PLoc.

Cell-PLoc: a package of Web servers for predicting subcellular localization of proteins in various organisms.

Motivation: With protein sequences entering into databanks at an explosive pace, the early determination of the family or subfamily class for a newly found enzyme molecule becomes important because this is directly related to the detailed information about which specific target it acts on, as well as to its catalytic process and biological function. Unfortunately, it is both time-consuming and costly to do so by experiments alone. In a previous study, the covariant-discriminant algorithm was introduced to identify the 16 subfamily classes of oxidoreductases. Although the results were quite encouraging, the entire prediction process was based on the amino acid composition alone without including any sequence-order information. Therefore, it is worthy of further investigation.

Results: To incorporate the sequence-order effects into the predictor, the 'amphiphilic pseudo amino acid composition' is introduced to represent the statistical sample of a protein. The novel representation contains 20 + 2λ discrete numbers: the first 20 numbers are the components of the conventional amino acid composition; the next 2λ numbers are a set of correlation factors that reflect different hydrophobicity and hydrophilicity distribution patterns along a protein chain. Based on such a concept and formulation scheme, a new predictor is developed. It is shown by the self-consistency test, jackknife test and independent dataset tests that the success rates obtained by the new predictor are all significantly higher than those by the previous predictors. The significant enhancement in success rates also implies that the distribution of hydrophobicity and hydrophilicity of the amino acid residues along a protein chain plays a very important role to its structure and function.

Contact: [email protected]

Kuo-Chen Chou

Papers

Prediction of protein cellular attributes using pseudo‐amino acid composition

Some remarks on protein attribute prediction and pseudo amino acid composition.

Prediction of protein structural classes.

Cell-PLoc: a package of Web servers for predicting subcellular localization of proteins in various organisms.

Using amphiphilic pseudo amino acid composition to predict enzyme subfamily classes