Forrest Sheng Bao's Cyberhome

I got quite confused about my compiler class this semester. The textbook we used is the famous "Dragoon Book" - Compilers, Principles, Techniques and Tools, by Alfred V. Aho, Monica S. Lam, Ravi Sethi and Jeffrey D. Ullman. Now I have figured everything out regarding syntax analysis. So I decide to write the algorithms in my way.

by Forrest Sheng Bao http://fsbao.net

In bioinformatics, we prefer constant or linear algorithms. The reason is obvious. But, sometimes, something strange just happened in some papers.

In 2006, Ho-Ming Chen, Yi-Hang Li and Shu-Hsing Wu published a paper on PNAS, "Bioinformatic prediction and experimental validation of a microRNA-directed tandem trans-acting siRNA cascade in Arabidopsis", http://www.pnas.org/content/104/9/3318.abstract.

by Forrest Sheng Bao http://fsbao.net

I think there is obvious a bug in Stellate Hamonie system, the software to sample data from EEG electrodes and manage them, convert them into EDF format. Actually, the EDF data converted from SIG has many problems, time stamps in wrong order, time stamps of wrong seconds, all-zero amplitudes, incorrect starting time, etc. I spent a lot of time to understand what's going on. At first, I was even frustrated that I couldn't use those data. Let's have a look at the data of their own SIG format and the data of EDF file converted from it.

by Forrest Sheng Bao http://fsbao.net

The 32-channel system (a.k.a. 10-20 system):

The 64-channel system:

The 128-channel system:

by Forrest Sheng Bao http://fsbao.net

Stellate is a company developing EEG sampling and analysis system in Montreal, Canada. Their system is good. But unlike many other software developed in Montreal Neurological Institute, McGill U., it runs on Windows! So, I tried to install it on Linux thru wine, another open source project aiming to run Windows application on Linux. Cool, it is very easy. Just 2 minutes work.

Here displays a short segment of 7 seconds in resolution 7.5 uV/mm.

by Forrest Sheng Bao http://fsbao.net

Some genes in DNA sequences are very short, for example 15 bases. They are too short to sequence. So, we can add some adapters at 5' and 3' to extend it to 34 bases. Besides, we have a huge pool of sequences from different sources and we mixed them up for sequencing. So we can also use the adapter as an indicator to mark the source of the gene.

by Forrest Sheng Bao http://fsbao.net

I am so excited today that I have found a very cool Python library for doing wavelet decomposition, the PyWavelets. It is developed by Filip Wasilewski

by Forrest Sheng Bao http://fsbao.net

data for healthy people

You can classify them by their brain activities: rest, sleep or cognitive activities.

And you can also classify them by the state of eyes: with eyes open or with eyes close.

by Forrest Sheng Bao http://fsbao.net

I tried to plot the data sourcing from a group at U. Bonn. http://www.epileptologie-bonn.de/front_content.php?idcat=193&lang=3&changelang=3

by Forrest Sheng Bao http://fsbao.net

I met a friend lately. He told me his group is using Nelder-Mead algorithm. I never heard about it, though I have used many optimization algorithms.

So, I googled it tonight and found it is really funny. But I doubt how efficient this algorithm is. For example, if the initial simplex is too small, you can't search much area.

So, maybe, if you really don't have an idea, grade search is not a bad option.

Here is a good online tutorial, one chapter of which is about optimization. http://www.boomer.org/c/p3/c11/c1106.html