|
< |
bimagicLab |
> |
|
bioimage informatics lab |
home \ people \ research \ publications
\ teaching \ contact
|
Biomolecular imaging The
question we would like to help answer is: What is the role and what can imaging
do for systems biology? In
recent years, the focus in biological sciences has shifted from understanding
single parts of larger systems, sort of vertical approach, to understanding complex
systems at the cellular and molecular levels, horizontal approach. Thus the
revolution of “omics” projects,
genomics and now proteomics. Understanding complexity of biological systems
is a task that requires acquisition, analysis and sharing of huge databases,
and in particular, high-dimensional databases. For example, in a project on
location proteomics, the fluorescence microscopy data sets can have a
dimension as high as 5: two spatial dimensions, z-stacks, time series and
different-color channels (different color probes for different proteins).
Processing such huge amount of bioimages visually
by biologists is inefficient, time-consuming and error-prone. Therefore, we
would like to move towards automated, efficient and robust processing of such
bioimage data sets. Moreover, some information hidden in the images may not
be readily visually available. For example, in the same project, two proteins
residing in the Golgi apparatus---giantin and
gpp130 cannot be distinguished better than randomly by humans, while when
employing data mining methods, they can be told apart. Therefore, we do not
only replace humans by machines for faster and more efficient processing but
also because new knowledge is generated through use of sophisticated
algorithms. The ultimate
dream is to have distributed but integrated large bioimage databases which
would allow researchers to upload their data, have it processed, share the
data, download data as well as platform-optimized code, etc, and all this in
a common format, something akin to the DICOM format for clinical imaging. To
achieve this goal, we must draw upon a whole host of sophisticated tools from
signal processing, machine learning and scientific computing. While such
tools are widely present in clinical (medical) imaging, they are not as
widespread in imaging of biological systems at cellular and molecular levels.
This is a huge challenge; the work below addresses a sample of existing
problems. |
|
Overview Sponsors Some of this material is
based upon work supported by the National Science Foundation under Grant No.
0515152 and 0331657 and the PA State Tobacco Settlement, Kamlet-Smith
Bioinformatics Grant. Any opinions, findings, and
conclusions or recommendations expressed in this material are those of the
author(s) and do not necessarily reflect the views of the sponsor(s). Collaborators PhD
students: Ramu Bhagavatula, Amina Chebira, Charles Jackson, Gowri Srinivasa MS
students: Thomas Merryman, Vivek Oak Undergraduate
students: Yann Barbotin, Lionel Coulot, Phil Cuadra, Manuel Gonzales-Rivero, Mukta Gore, Christina Hallock, Sarah Hsieh, Garrett Jenkinson,
Ryan Kellogg, Irina Khaimovich, Heather Kirschner, Abhay Mavalankar, Katie
Menzies, Alexia Mintos, Christina Onorato, Shauna Ormon, Keridon Williams Graduate
students: Sidharth Garg, Anupam Goyal, Yusong Guo,
Elvira Garcia Osuna Faculty
and staff: Carlos Castro, Justin Crowley, Matthew Fickus, Adam Linstedt, Jon
Minden, Jose Moura, Robert F. Murphy, Inci Özgüneş, John
Ozolek, Markus Püschel, Gustavo Rohde, Stefan Zappe Research
Corner Intelligent acquisition of fluorescence microscopy
images Segmentation in bioimaging Multiresolution classification in
bioimaging Tools for
multimodal neuroimaging data integration and
analysis Teaching
Corner Software Acquisition Segmentation Classification Multiresolution
classification |
|
Intelligent acquisition of fluorescence microscopy
images Adaptive
multiresolution acquisition of fluorescence microscopy data sets We propose an algorithm for
adaptive efficient acquisition of fluorescence microscopy data sets using a
multirate approach. We simulate acquisition as part of a larger system for
protein classification based on their subcellular location patterns, and thus
strive to maintain the achieved level of classification accuracy as much as
possible. This problem is similar to image compression but unique due to
additional restrictions, namely causality; we have access only to the
information that has been scanned up to that point. While we do want to
acquire fewer samples with as low distortion as possible to achieve
compression, our goal is to do so while affecting the overall classification
accuracy as little as possible. We achieve this by using an adaptive
multirate scanning scheme which samples the regions of the image area that
hold the most pertinent information. Our results show that we can achieve significant
compression which we can then use to acquire faster or to increase space
resolution of our data set, all while minimally affecting the classification
accuracy of the entire system. T.E.
Merryman and J. Kovačević, ''An adaptive multirate algorithm for acquisition of
fluorescence microscopy data sets'',
IEEE Trans. Image Proc., special issue on Molecular and Cellular Bioimaging,
September 2005. vol. 14, no. 9, September 2005, pp. 1246-1253. [rrc] T.E.
Merryman, J. Kovačević, E.G. Osuna and R.F. Murphy, ''Adaptive
multirate data acquisition of 3D cell images'', Proc. IEEE Int. Conf. Acoust., Speech, and Signal
Proc., Philadelphia, PA, March 2005, pp. II:133-136. Efficient acquisition and learning of
fluorescence microscopy data models We present a method to
efficiently acquire fluorescence microscopy datasets, to allow for higher
spatial and temporal resolution, and with less damage from photobleaching.
Our proposal is to restrict acquisition to regions where we expect to find an
object. Given that the objects are continuously moving, we must have an
accurate model to describe their motion to predict their future locations. We
outline a system for learning and applying this motion model, demonstrate its
application in a case study, and summarize results from more complex
applications. C.
Jackson, R.F. Murphy and J. Kovačević, "Efficient acquisition and learning of fluorescence
microscope data models", Proc. IEEE Conf. on
Image Proc., San Antonio, TX, Sep. 2007, pp. VI:245-248. |
|
Segmentation in bioimaging Topology
preserving STACS segmentation of protein subcellular location images We
present an algorithm for the segmentation of multicell
fluorescence microscopy images. Such images abound and a segmentation
algorithm robust to different experimental conditions as well as cell types
is becoming a necessity. In cellular imaging, among the most often used
segmentation algorithms is seeded watershed. One of its features is that it
tends to oversegment, splitting the cells, as well
as create segmented regions much larger than a true cell. This can be an advantage (the entire
cell is within the region) as well as a disadvantage (a large amount of
background noise is included). We present an algorithm which segments with
tight contours by building upon an active contour algorithm---STACS, by Pluempitiwiriyawej et al. We adapt the algorithm to suit
the needs of our data and use another technique, topology preservation by Han
et al., to build our topology preserving STACS (TPSTACS). Our algorithm
significantly outperforms the seeded watershed both visually as well as by
standard measures of segmentation quality: recall/precision, area similarity
and area overlap. L. Coulot, H. Kirschner, A. Chebira, J.M.F. Moura, J.
Kovačević, E.G. Osuna and R.F. Murphy, “Topology preserving STACS segmentation of protein
subcellular location images”, Proc. IEEE Intl. Symp. Biomed. Imaging, G. Srinivasa, V. Oak, S. Garg. M.
Fickus and J. Kovačević, "3D STACS segmentation of fMRI images of
the brain", Proc. IEEE Conf. on Image Proc., San Diego, CA, Oct.
2008. To appear. Multiscale
active contour segmentation In
recent years, the focus in biological science has shifted to understanding
complex systems at the cellular and molecular levels, a task greatly
facilitated by fluorescence microscopy. Segmentation, a fundamental yet
difficult problem, is often the first processing step following acquisition.
We have previously demonstrated that a stochastic active contour based
algorithm together with the concept of topology preservation (TPSTACS)
successfully segments single cells from multicell
images. In this paper we demonstrate that TPSTACS successfully segments
images from other imaging modalities such as DIC microscopy, MRI and fMRI.
While this method is a viable alternative to hand segmentation, it is not yet
ready to be used for high-throughput applications due to its large run time.
Thus, we highlight some of the benefits of combining TPSTACS with the multiresolution
approach for the segmentation of fluorescence microscope images. Here we
propose a multiscale active contour (MSAC)
transformation} framework for developing a family of modular algorithms for
the segmentation of fluorescence microscope images in particular, and
biomedical images in general. While this framework retains the flexibility
and the high quality of the segmentation provided by active contour-based
algorithms, it offers a boost in the efficiency as well as a framework to
compute new features that further enhance the segmentation. G. Srinivasa, M. C. Fickus and J. Kovačević, “Multiscale active contour transformations for the segmentation of fluorescence microscope images”, Proc. of SPIE Conf. on Wavelet Applications in Signal and Image Processing, San Diego, USA, Aug. 2007. Active
mask segmentation We
present a novel active mask framework for the segmentation of fluorescence
microscope images of cells, and in particular, for the segmentation of the
Golgi body as well as cell-volume computation. We demonstrate that the
algorithm is able to efficiently segment a stack of images and successfully
assign multiple pieces of the Golgi body in a 2D image to the cell to which
they belong. Further, we demonstrate that our algorithm is more accurate than
manual segmentation of these images.. G.
Srinivasa, M. Fickus, M. N.
Gonzales-Rivero, S. Y. Hsieh, Y. Guo, A. D. Linstedt and J. Kovačević,
“Active mask
segmentation for the cell-volume computation and Golgi-body segmentation of
HeLa cell images”, Proc. IEEE Intl. Symp. Biomed. Imaging, Paris,
France, May 2008, pp. 348-351. |
|
Multiresolution classification in bioimaging A multiresolution approach to automated classification
of protein subcellular location images Background: The problem of
automated interpretation of fluorescence microscope images depicting
subcellular protein locations is at the forefront of the current trend in
biology towards understanding the role and function of all proteins. Over the
past ten years, the feasibility of using machine learning methods to
recognize all major subcellular location patterns has been convincingly
demonstrated, using diverse feature sets and combinations of
classifiers. On a well-studied
data set of 2D HeLa single-cell images, the best performance to date, 91.5%,
was obtained upon the addition of a simple set of multiresolution features. Results: We report here a
novel approach for the classification of subcellular location patterns by classifying
in multiresolution subspaces. Our system is able to work with any feature set
and any classifier. It consists of multiresolution (MR) decomposition,
followed by feature computation and classification in each MR subspace,
yielding local decisions that are then combined into a global decision. With 26 texture features alone and a
neural network classifier, we obtained an increase in accuracy on the 2D HeLa
data set to 95.3%. Conclusions: We demonstrate
that the space-frequency localized information in the multiresolution
subspaces adds significantly to the discriminative power of the system.
Moreover, we show that a vastly reduced set of features is sufficient,
consisting of our novel modified Haralick texture
features. Our proposed system is general, allowing for any combinations of
sets of features and any combination of classifiers. A.
Chebira, Y. Barbotin, C. Jackson, T. Merryman, G. Srinivasa, R. F. Murphy and
J. Kovačević, “A multiresolution
approach to automated classification of protein subcellular location images,”
BMC Bioinformatics, vol. 8, no. 210, 2007. [rrc] G.
Srinivasa, T. Merryman, A. Chebira, A. Mintos and J. Kovačević,
“Adaptive
multiresolution techniques for subcellular protein location image
classification”, Proc. IEEE Int. Conf. Acoust., Speech, and Signal
Proc., Toulouse, France, May 2006, pp. V:1177-1180. Invited paper. T.
Merryman, K. Williams, G. Srinivasa, A. Chebira and J. Kovačević,
“A multiresolution enhancement to generic classifiers
of subcellular protein location images”, Proc. IEEE Intl.
Symp. Biomed. Imaging, Towards
an image analysis toolbox for high-throughput Drosophila embryo RNAi screens We build an image analysis
toolbox for high-throughput Drosophila
embryo RNAi screens. The goal is to tag the embryo as normal, developmentally
delayed or abnormal based on the ventral furrow formation. We break the problem into two parts:
in the first, we detect the developmental stage based on the progress of the
ventral furrow formation, and in the second, we tag the embryo as
normal/developmentally delayed/abnormal based on the stage detected and the
elapsed time. The crux of the algorithm is the multiresolution classifier,
and we show that, by classifying in multiresolution spaces, we obtain better
results than by classifying the embryo image alone. The final 2D accuracy
obtained was 93.17%, while by using 3D information, it increased to 98.35%. R. A.
Kellogg, A. Chebira, A. Goyal, P. A. Cuadra, S. F. Zappe, J. S. Minden and J.
Kovačević, “Towards an image analysis toolbox for high-throughput
Drosophila embryo RNAi screens”,
Proc. IEEE Intl. Symp. Biomed. Imaging, Classification in histopathology We propose a system for
identification of germ layer components in teratomas derived from human and
nonhuman primate embryonic stem cells.
Tissue regeneration and repair, drug testing and discovery, the cure
of genetic and developmental syndromes all may rest on the understanding of
the biology and behavior of embryonic stem (ES) cells. Within the field of stem cell biology,
an ES cell is not considered an ES cell until it can produce a teratoma tumor (the ``gold'' standard test); a seemingly
disorganized mass of tissue derived from all three embryonic germ layers;
ectoderm, mesoderm, and endoderm. Identification and quantification of tissue
types within teratomas derived from ES cells may expand our knowledge of
abnormal and normal developmental programming and the response of ES cells to
genetic manipulation and/or toxic exposures. In addition, because of the tissue
complexity, identifying and quantifying the tissue is tedious and time
consuming, but in turn the teratomas provides an excellent biological
platform to test robust image analysis algorithms. We use a multiresolution (MR)
classification system with texture features, as well as develop novel nuclear
texture features to recognize germ layer components. With redundant MR
transform, we achieve a classification accuracy of approximately 88%. Chebira,
J. A. Ozolek, C. A. Castro, W. G. Jenkinson, M. Gore, R. Bhagavatula, I.
Khaimovich, S. E. Ormon, C.
S. Navara, M. Sukhwani, K.
E. Orwig, A. Ben-Yehudah, G.
Schatten, G. K. Rohde and J.
Kovačević, “Multiresolution
identification of germ layer components in teratomas derived from human and
nonhuman primate embryonic stem cells”,
Proc. IEEE Intl. Symp. Biomed. Imaging, Paris, France, May 2008, pp. 979-982. Tools for multimodal neuroimaging
data integration and analysis We propose a novel method for
axonal bouton modeling and automated detection in populations of labeled
neurons, as well as bouton distribution analysis for the study of neural
circuit organization and plasticity.
Since axonal boutons are the presynaptic specializations of neural
synapses, their locations can be used to determine the organization of neural
circuitry, and in time-lapse studies, neural circuit dynamics. We propose simple geometric models for
axonal boutons that account for variations in size, position, rotation and
curvature of the axon in the vicinity of the bouton. We then use the normalized
cross-correlation between the model and image data as a test statistic for
bouton detection and position estimation. Thus, the problem is cast as a
statistical detection problem where we can tune the algorithm parameters to
maximize the probability of detection for a given probability of false
alarm. For example, we can detect
81% of boutons with 9% false alarm from noisy, out of focus, images. We also
present a novel method to characterize the orientation and elongation of a
distribution of labeled boutons and we demonstrate its performance by
applying it to a labeled data set. C.
A. Hallock, I.
Özgüneş, R. Bhagavatula, G. K. Rohde, J. C. Crowley, C. E. Onorato, A. Mavalankar, A. Chebira, C.
H. Tan and J.
Kovačević, “Axonal bouton modeling, detection and
distribution analysis for the study of neural circuit organization and
plasticity”, Proc. IEEE Intl. Symp. Biomed. Imaging, Paris, France, May
2008, pp. 165-168. |
|
C. Jackson, R.F. Murphy and
J. Kovačević, " Efficient acquisition and learning of
fluorescence microscopy data models", Proc. IEEE Conf. on Image Proc.,
San Antonio, TX, Sep. 2007. To appear. A.
Chebira, Y. Barbotin, C. Jackson, T. Merryman, G. Srinivasa, R. F. Murphy and
J. Kovačević, “A multiresolution approach to automated
classification of protein subcellular location images,” BMC
Bioinformatics, 2007. To appear. [rrc] R. A.
Kellogg, A. Chebira, A. Goyal, P. A. Cuadra, S. F. Zappe, J. S. Minden and J.
Kovačević, “Towards an image analysis toolbox for high-throughput
Drosophila embryo RNAi screens”,
Proc. IEEE Intl. Symp. Biomed. Imaging, T. Merryman, K. Williams, G. Srinivasa, A. Chebira and J.
Kovačević, “A
multiresolution enhancement to generic classifiers of subcellular protein
location images”,
Proc. IEEE Intl. Symp. Biomed. Imaging, L. Coulot, H. Kirschner, A. Chebira, J.M.F. Moura, J.
Kovačević, E.G. Osuna and R.F. Murphy, “Topology preserving STACS segmentation of protein
subcellular location images”, Proc. IEEE Intl. Symp. Biomed. Imaging, T.E.
Merryman and J. Kovačević, ''An adaptive multirate algorithm for acquisition of
fluorescence microscopy data sets'',
IEEE Trans. Image Proc., special issue on Molecular and Cellular Bioimaging,
September 2005. vol. 14, no. 9, September 2005, pp. 1246-1253. |
|
Next-generation bioimaging systems Efficient acquisition of fluorescence microscopy
data sets Topology preserving STACS segmentation of protein
subcellular location images Adaptive
multiresolution techniques for subcellular protein
location classification |
|
Center
for Bioimage Informatics (Kovačević & Murphy, MurphyLab (Murphy, Biomedical
Research Group
(Unser, EPFL) Danuser group (Scripps) Quantitative image analysis group (Olivo-Marin,
Pasteur) |