Discover biological features using composite images

• US 7,894,650 B2
• Filed: 11/13/2006
• Issued: 02/22/2011
• Est. Priority Date: 11/10/2005
• Status: Expired due to Fees

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Hunt, Brandon, Henle, Ernst S., Bondarenko, Andrey, Vega, Silvia C., Spiridonov, Alexander, Weng, Lee
• Primary Examiner(s): Le; Vu
• Assistant Examiner(s): Abdi; Amara
• Application Number:
  US11/599,184
• Publication Number:
  US 20070211928A1
• Time in Patent Office: 1,562 Days
• Field of Search: 382128-132, 250/281, 436/173
• US Class Current: 382/128
• CPC Class Codes: G01N 30/8624 Detection of slopes or peaks; baseline correction
• CPC Class Codes: G01N 30/8644 Data segmentation, e.g. time windows
• CPC Class Codes: G01N 30/8675 Evaluation, i.e. decoding of the signal into analytical information for analysis of specific compounds see also G01N30/88 and subgroups of G01N33/00; chemical libraries per se C40B
• CPC Class Codes: G01N 30/8679 Target compound analysis, i.e. whereby a limited number of peaks is analysed
• CPC Class Codes: G06T 2207/20212 Image combination
• CPC Class Codes: G06T 2207/30004 Biomedical image processing
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/41 based on statistical description of texture
• CPC Class Codes: G06V 10/40 Extraction of image or video features
• CPC Class Codes: G06V 10/7515 Shifting the patterns to accommodate for positional errors
• CPC Class Codes: G06V 20/695 Preprocessing, e.g. image segmentation
• CPC Class Codes: Y10T 436/24 Nuclear magnetic resonance, electron spin resonance or other spin effects or mass spectrometry

Abstract

An image processing system extracts parts or characteristics of interest from prepared biological samples One suitable use of the image processing system is to find biomarkers. But many other suitable uses are possible. Some components of the system include image preprocessing (data interpolation, retention time alignment, image noise filtering, background estimation, and formation of a composite image); image feature extraction (peaks, isotope groups, and charge groups); and computation of feature characteristics and expression statistics, differential expression, and non-differential expression. Outputs of the system include a candidate list of parts or characteristic of interest for aiding further discovery.

Claims

1. A system for discovering biological features, comprising:

a data collector configured to receive a set of images formed with chromatography instrument data and mass spectrometer data;

an image processor for processing a composite image, which is produced from the set of images, the image processing pipeline producing feature characteristics, expression profiles, and a candidate feature list; and

an image feature extractor is configured to identify isotope peaks from connected non-zero pixels on a composite image; and

split identified isotope peaks that are composed of two or more isotope peaks in a mass/charge direction, a retention time direction, or both.

2. The system of claim 1, wherein the feature characteristics provide information about a biological feature at three levels including peak, isotope group, and charge group.

3. The system of claim 2, wherein the peak is an isotopic peak defined by its contour at a specific retention time range and mass/charge range, wherein the isotope group is a group of isotopic peaks at the same charge state, and wherein the charge group is a collection of isotope groups.

4. The system of claim 3, wherein feature characteristics are selected from a group consisting of peak retention time start and end, peak mass/charge start and end, mass/charge centroid, charge state, and mass.

5. The system of claim 1, wherein expression profiles include a profile on peak intensity, which is a summation of intensity measurements of all non-zero pixels within a boundary contour of a peak.

6. The system of claim 1, wherein the candidate feature list is a set of peaks or isotope groups selected for protein identification.

7. A computer-implementable image processing pipeline, comprising:

a collector for collecting data from a process in which prepared biological samples are obtained from expression experiments under different treatment conditions;

an image processor for processing the data from the collector and forming a composite image; and

an image feature extractor for extracting feature characteristics from the composite image, which include peaks, isotope groups, and charge groups, wherein the image feature extractor is configured to identify isotope peaks from connected non-zero pixels on a composite image; and

split identified isotope peaks that are composed of two or more isotope peaks in a mass/charge direction, a retention time direction, or both.

8. The computer-implementable image processing pipeline of claim 7, wherein the image processor includes a retention time streak remover, a data interpolator, an image aligner, noise remover, and a background corrector.

9. The computer-implementable image processing pipeline of claim 7, further comprising an expression statistics processor for processing feature characteristics to produce expression profiles of replicates in all conditions at three summarization levels, which is selected from a group consisting of peak, isotope group, and charge group.

10. The computer-implementable image processing pipeline of claim 9, further comprising an expression analysis processor for processing expression profiles to produce a candidate list regarding differential and nondifferentially expressed features for biological target identifications.

11. The computer-implementable image processing pipeline of claim 7, wherein the image feature extractor is configured to compute isotope peak characteristics that are selected from a group consisting of peak mass/charge centroid, peak mass/charge width, peak time centroid, and peak time width.

12. The computer-implementable image processing pipeline of claim 7, wherein the image feature extractor is configured to identify isotope peaks that belong to an isotope group and labeling the isotope group with a unique isotope group index number.

13. The computer-implementable image processing pipeline of claim 12, wherein the image feature extractor is configured to compute isotope group characteristics that are selected from a group consisting of charge state, mono-isotopic mass/charge, and peptide mass.

14. The computer-implementable image processing pipeline of claim 13, wherein the image feature extractor is configured to process images of different treatment conditions to form a composite image;

and

find isotope peaks from connected pixels in the composite image that have intensity above a background noise parameter which is selected from a group consisting of the mean, median, maximum, minimum, and the standard deviation at a particular location in the composite image.

15. A computer-implementable image processing pipeline, comprising:

a collector for collecting data from a process in which prepared biological samples are obtained from expression experiments of different treatment conditions;

an image processor for processing the data from the collector and forming a composite image; and

an image feature extractor for extracting feature characteristics from the composite image, which include peaks, isotope groups, and charge groups, wherein the image feature extractor is configured to compute isotope group characteristics that are selected from a group consisting of charge state, mono-isotopic mass/charge, and peptide mass, and further wherein the image feature extractor is configured to identify an isotope group with only one isotope peak as belonging to a charge group with an unknown charge.

16. The computer-implementable image processing pipeline of claim 15, wherein the image feature extractor is configured to assign isotope groups having different charge states as belonging to one charge group if they have similar retention time and peptide mass.

17. A computer-implementable image processing pipeline, comprising:

a collector for collecting data from a process in which prepared biological samples are obtained from expression experiments of different treatment conditions;

an image processor for processing the data from the collector and forming a composite image; and

an image feature extractor for extracting feature characteristics from the composite image, which include peaks, isotope groups, and charge groups, wherein the image feature extractor is configured to remove streaks in a retention time direction by;

setting pixels, whose intensities rank in the lower ninety percentile, of a rasterized LC/MS image to zero to create a threshold LC/MS image;

creating an original bit mask;

performing a morphological open operation in the retention time direction on the original bit mask using a maximum peak width to reveal retention time streaks;

executing a conditional dilation operation to dilate the retention time streaks back to their original sizes in the retention time direction and in the mass/charge direction;

inverting the original bit mask to form an inverted bit mask; and

multiplying the inverted bit mask with the rasterized LC/MS image to remove the retention time streaks.