Orientation Assignment in Cryo-EM Amit Singer Princeton University Department of Mathematics and Program in Applied and Computational Mathematics July 24, 2014 Amit Singer (Princeton University) July 2014 1 / 30
Single Particle Reconstruction using cryo-EM Schematic drawing of the imaging process: The cryo-EM problem: Amit Singer (Princeton University) July 2014 2 / 30
Orientation Estimation: Fourier projection-slice theorem R i c ij c ij = ( x ij , y ij , 0) T ( x ij , y ij ) Projection I i ˆ Projection I i I i 3D Fourier space ( x ji , y ji ) Molecule φ R i c ij = R j c ji R i ∈ SO(3) ˆ Projection I j I j 3D Fourier space Electron source Cryo-EM inverse problem: Find φ (and R 1 , . . . , R n ) given I 1 , . . . , I n . n = 3: Vainshtein and Goncharov 1986, van Heel 1987 n > 3: S, Shkolnisky SIAM Imaging 2011 n > 3: Bandeira, Charikar, Chen, S, in preparation Amit Singer (Princeton University) July 2014 3 / 30
Assumptions for today’s talk Trivial point-group symmetry Homogeneity Amit Singer (Princeton University) July 2014 4 / 30
Experiments with simulated noisy projections Each projection is 129x129 pixels. SNR = Var( Signal ) Var( Noise ) , (b) SNR=2 0 (c) SNR=2 − 1 (d) SNR=2 − 2 (e) SNR=2 − 3 (a) Clean (f) SNR=2 − 4 (g) SNR=2 − 5 (h) SNR=2 − 6 (i) SNR=2 − 7 (j) SNR=2 − 8 Amit Singer (Princeton University) July 2014 5 / 30
Fraction of correctly identified common lines and the SNR Define common line as being correctly identified if both radial lines deviate by no more than 10 ◦ from true directions. Fraction p of correctly identified common lines increases by PCA log 2 (SNR) p 20 0.997 0 0.980 -1 0.956 -2 0.890 -3 0.764 -4 0.575 -5 0.345 -6 0.157 -7 0.064 -8 0.028 -9 0.019 Amit Singer (Princeton University) July 2014 6 / 30
Least Squares Approach R i c ij c ij = ( x ij , y ij , 0) T ( x ij , y ij ) ˆ Projection I i I i 3D Fourier space ( x ji , y ji ) R i c ij = R j c ji ˆ Projection I j I j 3D Fourier space Least squares: � � R i c ij − R j c ji � 2 min R 1 , R 2 ,..., R n ∈ SO (3) i � = j Search space is exponentially large and non-convex. Amit Singer (Princeton University) July 2014 7 / 30
Quadratic Optimization Under Orthogonality Constraints i � = j � R i c ij − R j c ji � 2 Quadratic cost: � Quadratic constraints: R T i R i = I 3 × 3 (det( R i ) = +1 constraint is ignored) We approximate the solution using SDP and rounding (S, Shkolnisky 2011) Related to: MAX-CUT (Goemans and Williamson 1995) Generalized Orthogonal Procrustes Problem (Nemirovski 2007) PhaseLift (Candes, Strohmer, Voroninski 2013) Non-commutative Grothendick Problem (Naor, Regev, Vidick 2013) “Robust” version – Least Unsquared Deviations (Wang, S, Wen 2013) � min � R i c ij − R j c ji � R 1 , R 2 ,..., R n ∈ SO (3) i � = j Amit Singer (Princeton University) July 2014 8 / 30
Detour: MAX-CUT MAX-CUT: Given a weighted undirected graph with n vertices and non-negative weights w ij ≥ 0, split the vertices into two sets such that the total weight of edges across the cut is maximized. 1 w 14 w 12 4 w 24 2 w 45 w 25 w 34 5 3 w 35 CUT( { 1 , 2 , 3 } , { 4 , 5 } ) = w 14 + w 24 + w 25 + w 34 + w 35 Amit Singer (Princeton University) July 2014 9 / 30
The MAX-CUT Approximation of Goemans & Williamson 1 � OPT = max w ij (1 − y i y j ) 4 y 1 , y 2 ,..., y n ∈{− 1 , +1 } i � = j Combinatorial problem that is known to be NP-complete SDP relaxation of Goemans-Williamson (1995): The n × n Gram matrix G = yy T ( G ij = y i y j ) is PSD ( G � 0), G ii = 1, rank( G ) = 1. 1 w ij − 1 � max 4 Tr( GW ) 4 G ∈ R n × n i , j s.t. G � 0 , G ii = 1 , i = 1 , 2 , . . . , n . Randomize a random vector z with i.i.d standard Gaussian entries z i ∼ N (0 , 1), and compute Yz , where G = YY T is the Cholesky decomposition of G . Round y i = sign[( Yz ) i ]. Theorem (G-W): E [CUT GW ] > 0 . 87 · OPT. Amit Singer (Princeton University) July 2014 10 / 30
SDP Relaxation for the Common-Lines Problem Least squares is equivalent to maximizing the sum of inner products: � R i c ij − R j c ji � 2 ⇐ � � min ⇒ max � R i c ij , R j c ji � R 1 , R 2 ,..., R n ∈ SO (3) R 1 , R 2 ,..., R n ∈ SO (3) i � = j i � = j � Tr( c ji c T ij R T ⇐ ⇒ i R j ) ⇐ ⇒ max R 1 , R 2 ,..., R n ∈ SO (3) Tr( CG ) max R 1 , R 2 ,..., R n ∈ SO (3) i � = j C is the 2 n × 2 n matrix (“the common lines matrix”) with � 0 � x ji � x ji x ij � � � � x ji y ij 0 c T � C ij = ˜ c ji ˜ ij = x ij y ij = , C ii = y ji y ji x ij y ji y ij 0 0 R T ˜ G is the 2 n × 2 n Gram matrix G = ˜ R with G ij = ˜ i ˜ R T R j : ˜ ˜ 1 ˜ ˜ 1 ˜ R T R T R T · · · I 2 × 2 R 2 R n 1 ˜ ˜ 2 ˜ ˜ 2 ˜ R T R T R T · · · � ˜ R 1 I 2 × 2 R n 2 ˜ ˜ � G = R 1 R 2 · · · R n = . . . . ... . . . . . . . . ˜ ˜ n ˜ ˜ n ˜ R T R T R T · · · R 1 R 2 I 2 × 2 n Amit Singer (Princeton University) July 2014 11 / 30
SDP Relaxation and Rounding R 1 , R 2 ,..., R n ∈ SO (3) Tr( CG ) max SDP Relaxation : G ∈ R 2 n × 2 n Tr( CG ) max s.t. G � 0 , G ii = I 2 × 2 , i = 1 , 2 , . . . , n . Missing is the non-convex constraint rank( G ) = 3. Randomize a 2 n × 3 orthogonal matrix Q using (careful) QR factorization of a 2 n × 3 matrix with i.i.d standard Gaussian entries Compute Cholesky decomposition G = YY T ⇒ ˜ Round using SVD: ( YQ ) i = U i Σ i V T R T = U i V T = i . i i Use the cross product to find R T i . Loss of handedness. Amit Singer (Princeton University) July 2014 12 / 30
Spectral Relaxation for Uniformly Distributed Rotations � x 1 x 2 � | | � � i i R 1 R 2 y 1 y 2 = i = 1 , . . . , n . , i i i i z 1 z 2 | | i i Define 3 vectors of length 2 n � T � x 1 x 2 x 1 x 2 x 1 x 2 · · · x = 1 1 2 2 n n � T � y 1 y 2 y 1 y 2 y 1 y 2 = · · · y 1 1 2 2 n n � T � z 1 z 2 z 1 z 2 z 1 z 2 = · · · z 1 1 2 2 n n Rewrite the least squares objective function as R 1 ,..., R n ∈ SO (3) x T Cx + y T Cy + z T Cz � � R i c ij , R j c ji � = max max R 1 ,..., R n ∈ SO (3) i � = j By symmetry , if rotations are uniformly distributed over SO (3), then the top eigenvalue of C has multiplicity 3 and corresponding eigenvectors are x , y , z from which we recover R 1 , R 2 , . . . , R n ! Amit Singer (Princeton University) July 2014 13 / 30
Spectrum of C Numerical simulation with n = 1000 rotations sampled from the Haar measure; no noise. Bar plot of positive (left) and negative (right) eigenvalues of C : 600 180 160 500 140 400 120 100 − λ λ 300 80 200 60 40 100 20 0 0 0 10 20 30 40 50 60 0 10 20 30 40 50 60 Eigenvalues: λ l ≈ n ( − 1) l +1 l = 1 , 2 , 3 , . . . . ( 1 2 , − 1 6 , 1 l ( l +1) , 12 , . . . ) Multiplicities: 2 l + 1. Two basic questions: Why this spectrum? Answer: Representation Theory of SO(3) 1 (Hadani, S, 2011) Is it stable to noise? Answer: Yes, due to random matrix theory. 2 Amit Singer (Princeton University) July 2014 14 / 30
Common Lines Kernel R 3 i × R 3 j Common line equation: R i c ij = R j c ji = ± � R 3 i × R 3 j � R 3 i × R 3 R 3 i × R 3 j j c ij = ± R − 1 c ji = ± R − 1 j � , i � R 3 i × R 3 j � R 3 i × R 3 j � Basic properties of the cross product: C ij is only a function of the ratio U ij = R − 1 R j i � x ij x ji � x ij � � � x ij y ji = ( Pc ij )( Pc ji ) T � = = C ij x ji y ji y ij x ji y ij y ji y ij � 1 [ I 3 × U 3 ij ) 3 × I 3 ] T ij ][( U − 1 � 0 0 P T , where P = = P . � I 3 × U 3 ij � 2 0 1 0 R j ) : R 2 �→ R 2 . Common lines kernel: C ij = k ( R i , R j ) = ˜ k ( R − 1 i Amit Singer (Princeton University) July 2014 15 / 30
Common Lines Operator R j ) : R 2 �→ R 2 . Common lines kernel: C ij = k ( R i , R j ) = ˜ k ( R − 1 i Suppose f : SO (3) �→ R 2 then n n ( Cf )( R i ) = 1 1 � � C ij f ( R j ) ≈ k ( R i , R j ) f ( R j ) d µ ( R j ) = ( C f )( R i ) . n SO (3) j =1 Common lines integral operator C is a convolution over SO(3): � ˜ k ( R − 1 ( C f )( R i ) = R j ) f ( R j ) d µ ( R j ) = λ f ( R i ) i SO (3) Representation theory of SO(3) explains eigenvalues and eigenfunctions. Eigenfunctions are tangent vector fields to the 2-sphere (sections of the tangent bundle). They are the projected gradient of the scalar spherical harmonics. Amit Singer (Princeton University) July 2014 16 / 30
Probabilistic Model and Wigner’s Semi-Circle Law Simplistic Model: every common line is detected correctly with probability p , independently of all other common-lines, and with probability 1 − p the common lines are falsely detected and are uniformly distributed over the unit circle. Let C clean be the matrix C when all common-lines are detected correctly ( p = 1). The expected value of the noisy matrix C is E [ C ] = pC clean , as the contribution of the falsely detected common lines to the expected value vanishes . Decompose C as C = pC clean + W , where W is a 2 n × 2 n zero-mean random matrix. The eigenvalues of W are distributed according to Wigner’s semi-circle law whose support, up to O ( p ) and finite sample √ √ fluctuations, is [ − 2 n , 2 n ]. Amit Singer (Princeton University) July 2014 17 / 30
Recommend
More recommend