Mesa Analytics : Home

Welcome

Mesa Analytics & Computing, Inc. provides consulting and creates custom data mining solutions and software for companies with problems in knowledge discovery, dynamic web applications for simulation software companies, and compound acquisition decisions and drug design for pharmaceutical and biotech companies.

Mesa Analytics & Computing, Inc.

Check out our book due out this summer written by John and Norah MacCuish: Clustering in Bioinformatics and Drug Discovery

Mesa Analytics has a Molecular Shape Fingerprinter using quasi-Monte Carlo Integration‏ Please contact us if you would like to become one of the beta testing sites.

When you turn to Mesa Analytics & Computing, you receive the full benefit of our expertise. We work with you to design custom solutions in knowledge discovery and data mining, expertise in high performance computing and software, and custom dynamic web application systems.

Expertise in mining data, includes:

High-throughput screening
Gene expression
Database systems
Control systems
Credit card products
Precision agriculture
Market analysis
Compound acquisition decisions

We employ state-of-the-art tools and techniques in areas such as:

Multivariate analysis
Statistical modeling
Cluster analysis
Anomaly detection
Market basket analysis
Chemical diversity analysis
Object relational systems
Time series
Cheminformatics
Compound Acquisition decisions

If your company falls into one of the above categories or has a problem similar to those given above, please contact us.

Dynamic Web Applications

Mesa Analytics has extensive experience in web application development. We have developed dynamic web applications using Ajax techniques to support cheminformatics learning; cheminformatics workflows for drug discovery; analysis of pandemic mitigation simulations; asset life-cycle cost analysis for the power utility industry; and video-centric collaborative scientific research.

Our web applications use a variety of frameworks, including Django, CherryPy, and ASP.NET; client-side toolkits such as jQuery, jQuery UI and Flex 3 and DBMS systems including PostgreSQL, MySQL, MS SQL Server and SQLite.

Our web applications typically provide custom, interactive data visualizations. We have used jMol and ChemAxon’s Marvin tools to present 3D chemical structure information. We have developed our own custom 2D depictions using both HTML5 Canvas and ActionScript. For statistical summary visualizations we often use Gnu R, running in background processes which are managed by our web servers. Finally, we have built custom, interactive 2D scatterplots using both Processing and ActionScript.

H1N1 'Swine Flu’ Incident Dynamic Map – Select the Map to interact with current HIN1 data (zoom and slider bar)

Mesa Analytics has created a prototype “solar installation site analysis tool”

Custom Data Mining Software

For customers who require capabilities not found in our commercial data mining tools, Mesa Analytics develops custom data mining software.

Fingerprint Module

320 MDL keys 2D fingerprinting – As published by MDL in JCICS
Uses OEChem from OpenEye Scientific Software
Fingerprint Theory Manual
“Substructure commonality analysis and visualization with new key-based binary fingerprinter”, Norah MacCuish and John MacCuish, ACS National Spring National Meeting in Chicago, CINF Session, March 24-28, 2007

Grouping Module

Measures:
- Tversky
- Tanimoto
- Euclidean
- Ochiai (1 – Cosine)
- Hamman
- Euclidean*Soergel product
- Euclidean*Cosine
Grouping:
- Taylor’s – symmetric and asymmetric, in disjoint and non-disjoint (overlapping) versions
  - MPI / PVM enabled
- Hierarchical (RNN Wards, Complete Link, Group Average)
- Kmeans/Kmodal – disjoint and non-disjoint
- Jarvis Patrick – disjoint and non-disjoint
Ambiguity Statistics – pick thresholds for the least ambiguous results

The tools in the grouping module can use dissimilarity matrices generated via non-binary measures and descriptors. They can be used with other applications which produce binary data or real valued data.

For example, presently, an academic user is using the asymmetric clustering tools for analysis of Financial Service Industry data. (See: Incorporating sequential information into traditional classification models by using an element/position-sensitive SAM, Anita Prinzie and Dirk Van den Poel, April, 2005. Available online at Decision Support Systems.)

For more information see the Grouping Module Manual.

ChemTattoo Module

Generate Modal (“Common Feature”) Fingerprints
Modal Fingerprint Statistics
Common features, MACCS key counts, and atom counts are identified, with bond and atom coloring depiction information output
Depict the structural commonalities of compound sets through Openeye’s Ogham or ChemAxon’s Marvin Tools

SAESAR

SAESAR is a desktop application for exploring relationships between empirical properties and shape, electrostatics and color.

Please contact us if you would like an evaluation copy.

"If in your science you only look for business, then you risk finding neither knowledge nor business."

— Haldor Topsøe