Secondary Structure Prediction Lecture 7 Structural Bioinformatics Dr

Secondary Structure Prediction Lecture 7 Structural Bioinformatics Dr. Avraham Samson 81 -871

Secondary structure prediction from amino acid sequence 2012 Avraham Samson - Faculty of Medicine - Bar Ilan University 2

Secondary Structure Prediction • Given a protein sequence a 1 a 2…a. N, secondary structure prediction aims at defining the state of each amino acid ai as being either H (helix), E (extended=strand), or O (other) (Some methods have 4 states: H, E, T for turns, and O for other). • The quality of secondary structure prediction is measured with a “ 3 -state accuracy” score, or Q 3 is the percent of residues that match “reality” (X-ray structure).

Quality of Secondary Structure Prediction Determine Secondary Structure positions in known protein structures using DSSP or STRIDE: 1. Kabsch and Sander. Dictionary of Secondary Structure in Proteins: pattern recognition of hydrogen-bonded and geometrical features. Biopolymer 22: 2571 -2637 (1983) (DSSP) 2. Frischman and Argos. Knowledge-based secondary structure assignments. Proteins, 23: 566 -571 (1995) (STRIDE)

Limitations of Q 3 ALHEASGPSVILFGSDVTVPPASNAEQAK hhhhhooooeeeeooooohhhhh Amino acid sequence ohhhooooeeeeoooooeeeooohhhhhh Q 3=22/29=76% Actual Secondary Structure (useful prediction) hhhhhoooohhhhooooohhhhh Q 3=22/29=76% (terrible prediction) Q 3 for random prediction is 33% Secondary structure assignment in real proteins is uncertain to about 10%; Therefore, a “perfect” prediction would have Q 3=90%.

Early methods for Secondary Structure Prediction • Chou and Fasman (Chou and Fasman. Prediction of protein conformation. Biochemistry, 13: 211 -245, 1974) • GOR (Garnier, Osguthorpe and Robson. Analysis of the accuracy and implications of simple methods for predicting the secondary structure of globular proteins. J. Mol. Biol. , 120: 97120, 1978)

Chou and Fasman • Start by computing amino acids propensities to belong to a given type of secondary structure: Propensities > 1 mean that the residue type I is likely to be found in the Corresponding secondary structure type.

Chou and Fasman Amino Acid Ala Cys Leu Met Glu Gln His Lys Val Ile Phe Tyr Trp Thr Gly Ser Asp Asn Pro Arg -Helix 1. 29 1. 11 1. 30 1. 47 1. 44 1. 27 1. 22 1. 23 0. 91 0. 97 1. 07 0. 72 0. 99 0. 82 0. 56 0. 82 1. 04 0. 90 0. 52 0. 96 -Sheet 0. 90 0. 74 1. 02 0. 97 0. 75 0. 80 1. 08 0. 77 1. 49 1. 45 1. 32 1. 25 1. 14 1. 21 0. 92 0. 95 0. 72 0. 76 0. 64 0. 99 Turn 0. 78 0. 80 0. 59 0. 39 1. 00 0. 97 0. 69 0. 96 0. 47 0. 51 0. 58 1. 05 0. 75 1. 03 1. 64 1. 33 1. 41 1. 23 1. 91 0. 88 Favors -Helix Favors -strand Favors turn

Chou and Fasman Predicting helices: - find nucleation site: 4 out of 6 contiguous residues with P( )>1 - extension: extend helix in both directions until a set of 4 contiguous residues has an average P( ) < 1 (breaker) - if average P( ) over whole region is >1, it is predicted to be helical Predicting strands: - find nucleation site: 3 out of 5 contiguous residues with P( )>1 - extension: extend strand in both directions until a set of 4 contiguous residues has an average P( ) < 1 (breaker) - if average P( ) over whole region is >1, it is predicted to be a strand

Chou and Fasman f(i) Position-specific parameters for turn: Each position has distinct amino acid preferences. Examples: -At position 2, Pro is highly preferred; Trp is disfavored -At position 3, Asp, Asn and Gly are preferred -At position 4, Trp, Gly and Cys preferred f(i+1) f(i+2) f(i+3)

Chou and Fasman Predicting turns: - for each tetrapeptide starting at residue i, compute: - PTurn (average propensity over all 4 residues) - F = f(i)*f(i+1)*f(i+2)*f(i+3) - if PTurn > P and PTurn > 1 and F>0. 000075 tetrapeptide is considered a turn. Chou and Fasman prediction: http: //fasta. bioch. virginia. edu/fasta_www/chofas. htm

The GOR method Position-dependent propensities for helix, sheet or turn is calculated for each amino acid. For each position j in the sequence, eight residues on either side are considered. j A helix propensity table contains information about propensity for residues at 17 positions when the conformation of residue j is helical. The helix propensity tables have 20 x 17 entries. Build similar tables for strands and turns. GOR simplification: The predicted state of AAj is calculated as the sum of the positiondependent propensities of all residues around AAj. GOR can be used at : http: //abs. cit. nih. gov/gor/ (current version is GOR IV)

Accuracy • Both Chou and Fasman and GOR have been assessed and their accuracy is estimated to be Q 3=60 -65%. (initially, higher scores were reported, but the experiments set to measure Q 3 were flawed, as the test cases included proteins used to derive the propensities!)

Neural networks The most successful methods for predicting secondary structure are based on neural networks. The overall idea is that neural networks can be trained to recognize amino acid patterns in known secondary structure units, and to use these patterns to distinguish between the different types of secondary structure. Neural networks classify “input vectors” or “examples” into categories (2 or more). They are loosely based on biological neurons.

The perceptron X 1 w 1 T X 2 XN Input w 2 w. N Threshold Unit Output The perceptron classifies the input vector X into two categories. If the weights and threshold T are not known in advance, the perceptron must be trained. Ideally, the perceptron must be trained to return the correct answer on all training examples, and perform well on examples it has never seen. The training set must contain both type of data (i. e. with “ 1” and “ 0” output).

Applications of Artificial Neural Networks • • speech recognition medical diagnosis image compression financial prediction

Existing Neural Network Systems for Secondary Structure Prediction • First systems were about 62% accurate. • Newer ones are about 70% accurate when they take advantage of information from multiple sequence alignment. • PHD • NNPREDICT

Applications in Bioinformatics • Translational initiation sites and promoter sites in E. coli • Splice junctions • Specific structural features in proteins such as α-helical transmembrane domains

Neural Networks Applied to Secondary Structure Prediction • Create a neural network (a computer program) • “Train” it uses proteins with known secondary structure. • Then give it new proteins with unknown structure and determine their structure with the neural network. • Look to see if the prediction of a series of residues makes sense from a biological point of view – e. g. , you need at least 4 amino acids in a row for an αhelix.

Example Neural Network Training pattern One of n inputs, each with 21 bits From Bioinformatics by David W. Mount, p. 453

Inputs to the Network • Both the residues and target classes are encoded in unary format, for example • Alanine: 1 0 0 0 0 0 • Cysteine: 0 1 0 0 0 0 0 • Helix: 1 0 0 • Each pattern presented to the network requires n 21 -bit inputs for a window of size n. (One bit is required per residue to indicate when the window overlaps the end of the chain). • The advantage of this sparse encoding scheme is that it does not pay attention to ordering of the amino acids • The main disadvantage is that it requires a lot of input.

Weights • Input values at each layer are multiplied by weights. • Weights are initially random. • Weights are adjusted after the output is computed based on how close the output is to the “right” answer. • When the full training session is completed, the weights have settled on certain values. • These weights are then used to compute output for new problems that weren’t part of the training set.

Neural Network Training Set • A problem-solving paradigm modeled after the physiological functioning of the human brain. • A typical training set contains over 100 nonhomologous protein chains comprising more than 15, 000 training patterns. • The number of training patterns is equal to the total number of residues in the 100 proteins. • For example, if there are 100 proteins and 150 residues per protein there would be 15, 000 training patterns.

Neural Network Architecture • A typical architecture has a window-size of n and 5 hidden layer nodes. * • Then a fully-connected would be 17(21)-5 -3 network, i. e. a net with an input window of 17, five hidden nodes in a single hidden layer and three outputs. • Such a network has 357 input nodes and 1, 808 weights. • ((17 * 21) * 5) + (5 * 3) + 5 + 3 = 1808? *This information is adapted from “Protein Secondary Structure Prediction with Neural Networks: A Tutorial” by Adrian Shepherd (UCL), http: //www. biochem. ucl. ac. uk/~shepherd/sspred_tutorial/ss-index. html. )

Window • The n-residue window is moved across the protein, one residue at a time. • Each time the window is moved, the center residue becomes the focus. • The neural network “learns” what secondary structure that residue is a part of. It keeps adjusting weights until it gets the right answer within a certain tolerance. Then the window is moved to the right.

Predictions Based on Output • Predictions are made on a winner-takes-all basis. • That is, the prediction is determined by the strongest of the three outputs. For example, the output (0. 3, 0. 1) is interpreted as a helix prediction.

Disadvantages to Neural Networks • They are black boxes. They cannot explain why a given pattern has been classified as x rather than y. Unless we associate other methods with them, they don’t tell us anything about underlying principles.

Summary • Perceptrons (single-layer neural networks) can be used to find protein secondard structure, but more often feed-forward multi-layer networks are used. • Two frequently-used web sites for neuralnetwork-based secondary structure prediction are PHD (http: //www. emblheidelberg. de/predictprotein. html ) and NNPREDICT (http: //www. cmpharm. ucsf. edu/~nomi/nnpredict. html)

The perceptron Notes: - The input is a vector X and the weights can be stored in another vector W. - the perceptron computes the dot product S = X. W - the output F is a function of S: it is often set discrete (i. e. 1 or 0), in which case the function is the step function. For continuous output, often use a sigmoid: 1 1/2 0 0 - Not all perceptrons can be trained ! (famous example: XOR)

The perceptron Training a perceptron: Find the weights W that minimizes the error function: P: number of training data Xi: training vectors F(W. Xi): output of the perceptron t(Xi) : target value for Xi Use steepest descent: - compute gradient: - update weight vector: - iterate (e: learning rate)

Neural Network A complete neural network is a set of perceptrons interconnected such that the outputs of some units becomes the inputs of other units. Many topologies are possible! Neural networks are trained just like perceptron, by minimizing an error function:

Neural networks and Secondary Structure prediction Experience from Chou and Fasman and GOR has shown that: – In predicting the conformation of a residue, it is important to consider a window around it. – Helices and strands occur in stretches – It is important to consider multiple sequences

PHD: Secondary structure prediction using NN

PHD: Input For each residue, consider a window of size 13: 13 x 20=260 values

PHD: Network 1 Sequence Structure 13 x 20 values 3 values Network 1 P (i) Pc(i)

PHD: Network 2 Structure 3 values For each residue, consider a window of size 17: 3 values 17 x 3=51 values Network 2 P (i) Pc(i)

PHD • Sequence-Structure network: for each amino acid aj, a window of 13 residues aj-6…aj…aj+6 is considered. The corresponding rows of the sequence profile are fed into the neural network, and the output is 3 probabilities for aj: P(aj, alpha), P(aj, beta) and P(aj, other) • Structure-Structure network: For each aj, PHD considers now a window of 17 residues; the probabilities P(ak, alpha), P(ak, beta) and P(ak, other) for k in [j-8, j+8] are fed into the second layer neural network, which again produces probabilities that residue aj is in each of the 3 possible conformation • Jury system: PHD has trained several neural networks with different training sets; all neural networks are applied to the test sequence, and results are averaged • Prediction: For each position, the secondary structure with the highest average score is output as the prediction

PSIPRED Jones. Protein secondary structure prediction based on position specific scoring matrices. J. Mol. Biol. 292: 195 -202 (1999) Convert to [0 -1] Using: Add one value per row to indicate if Nter of Cter

Performances (monitored at CASP) CASP 1 YEAR 1994 # of Targets 6 <Q 3> Group 63 Rost and Sander Rost CASP 2 1996 24 70 CASP 3 1998 18 75 Jones CASP 4 2000 28 80 Jones

Secondary Structure Prediction -Available servers: - JPRED : http: //www. compbio. dundee. ac. uk/~www-jpred/ - PHD: http: //cubic. bioc. columbia. edu/predictprotein/ - PSIPRED: http: //bioinf. cs. ucl. ac. uk/psipred/ - NNPREDICT: http: //www. cmpharm. ucsf. edu/~nomi/nnpredict. html - Chou and Fassman: http: //fasta. bioch. virginia. edu/fasta_www/chofas. htm -Interesting paper: - Rost and Eyrich. EVA: Large-scale analysis of secondary structure prediction. Proteins 5: 192 -199 (2001)

Solvent accessibility prediction http: //140. 113. 239. 214/~weilun/ Membrane region prediction http: //www. sbc. su. se/~miklos/DAS 2012 Avraham Samson - Faculty of Medicine - Bar Ilan University 41