COMPASS A.I. Software
Magellan's machine learning tools help realize the untapped potential of Mass Spec-based OMICs
Thanks to COMPASSS AI software, Mass Spec-based research programs can rapidly extract insights from their data and quickly move to the next study
​
COMPASS uses the entirety of your mass spec data, so no finding is missed and confidence in your results is fully supported
COMPASS enhances discovery…
-
Rapidly extracts ALL of the data connected to biological differences
-
Works with even the largest datasets
-
Identifies key molecules without bias towards abundance or familiarity
-
Provides built-in visual outputs that facilitate biological insight
-
Compares signal and noise for confidence that goes beyond standard statistics
…and has many research applications
-
Define molecular differences between the most subtle biological differences
-
Track molecular changes across time or across populations
-
Build OMICs-based machine learning tests for biological differences
-
Use machine learning to stratify subjects based on mass spec-based OMICs
COMPASS AI: Introductory Videos
Redefining data scale limits for mass spec-based OMICs
Preventing gaps in Mass Spec-Based OMICs datasets
Avoiding data loss or reduction in Mass Spec-Based OMICs
The Problem with Current Mass Spec Software
PROBLEM: Mass spec data files are large and intractable to rapid analysis at scale ​
-
Huge file sizes and proprietary, inaccessible formats
-
Highly technical, focused software solutions
-
Software does not keep pace with instrument output
-
Workflows designed to generate results that are easy to consume, not for maximum insight
​
THE STATUS QUO: Mass Spec-based OMICs studies are limited in scope
-
Study sizes are highly constrained
-
Studies ignore information-rich data
-
Data loss and data gaps are accepted
-
Data structure is not tractable for modern A.I. tools
-
Inability to cross-compare data from different studies
​
Use COMPASS to bring your mass spec research into the information age
Solution: COMPASS
Unbiased
Considers data for all the molecular species instrument identifies
Works for different study designs-- proteomics, lipidomics, and metabolomics, and beyond
Agnostic to instrument and original output data file type
Can work with data (MS1) that are ignored in most other analyses
Comprehensive
Considers the data in its entirety, not a selection only
Total integration of mass spec datasets for parallel analysis of proteomics, lipidomics, and metabolomics data
Finds the most biologically relevant information, even if unexpected
Assesses validity of data signals across entire dataset
Fast
Turns data into insight in hours, not weeks or months, with visual outputs that allow for easy interpretation of results
​
Determine if a study generates discernable differences at the outset of analysis
Perform analyses using different parameters to maximize information extraction
Rapidly adjust to the data by altering experimental design or moving to the next study
Powerful
Uses proprietary computational tools to extract maximum information
Data restructuring is built for machine learning and AI
Designed for analysis of the most complex samples—efforts to reduce sample complexity are unnecessary
Designed for large studies—assess differences in the largest studies