I'm involved in a research on predicting the Membrane Spanning Regions (MSRs) of Trans Membrane Proteins (TMPs) using HMM. If you are interested in this topic, please feel free to publish your comments....
Asela TAK.
Thank You.
Subscribe to:
Post Comments (Atom)
7 comments:
this is a interesting topic
wow..wonderful. I am interested to know more about your research in detail. My MS project is also in bioinformatics. Hopefully this will be a good place to exchange our knowledge.
-Lankara
Basically, my research exploits the similarity between speech signals and the characteristics of TMPs. HMMs are used to model these unique characteristics of TMPs. My aim is to increase the accuracy of MSR predictions than currently available HMM based prediction methods.
Hi Everybody,
I have completed this research and achieved interesting MSR prediction results.
Used new type of HMM to model the segments of TMPs. Its characteristics are parallel to biological systems such as randomness and recurring nature.
When benchmarked my proposed method is the best method for predicting MSR comprehensively, having almost 20% more comprehensiveness than best existing method. However, it is in 8th place when it comes to accuracy of MSR prediction.
For results to be more accurate I have removed the homology relationships between training and testing TMP sets, which has put the prediction method into rigorous test.
Asela Karunarathna
"Bioinformatics"..
If you can use full terminology instead of abbreviations in the first discussion it will be useful for a person like me to understand more about the topic. I am not directly involved in membrane modeling, but I work on the application (engineering) side of membranes and membrane proteins...But I would like to follow this blog and have my comments when I feel useful.
Thanks
Asela - You know me!
The acronyms go as follows,
TMP = Trans-Membrane Proteins (Proteins that span through and around the phospholipids bi-layer membrane), an integral protein of membrane
MSR = Membrane Spanning Regions (Segment of the TMP that goes through / spans the membrane, this segment is water fearing / hydrophobic)
HMM = Hidden Markov Model (State is not directly observable in HMM, though parameters / output influenced by state are visible)
So ultimately what this research has done is predicting the locations of MSRs along the TMP. Have achieved this by modeling different unique segments of TMP using HMM that was customized for biological systems.
Asela
matanam Greek wage :)
sounds like you have achieved what you wanted so welldone!
Menol (From virtusa)
Post a Comment