Hi,
I construct a profile HMM with randomized uniform distributions. I then train
the model and, regardless of the training set I use, it ends up looking like
this:
- a 0.999 probability to go from one delete state to the next
- roughly a uniform emission distribution for emission states (insert
Hi,
I am aligning a set of sequences to a profile HMM. I would like to convert the
individual alignments into one (multiple sequence) alignment object. I tried to
build a new map for my msa by retrieving the sequence symbollist from the
statepaths and the sequence names. The problem with this a
Hi,
If I create a uniform protein distribution and check the emission
probabilities, I find that they are set to 1/21 rather than 1/20...
Henry Romijn
FiniteAlphabet emissionAlpha =
(FiniteAlphabet) AlphabetManager.alphabetForName("PROTEIN");
ProfileHMM profile = new ProfileHMM(
Hi,
I have narrowed down the problem in creating a dp object for a profile HMM to a
possible bug in the code below from the DP class. I could really use some help
here..
Thanks,
Henry Romijn
public static State[] stateList(MarkovModel mm)
throws IllegalSymbolException, IllegalTransitionE
Hi,
After I have successfully created and initialized a profile HMM, I try to
register the model with a new DP object as follows:
DP dp = DPFactory.DEFAULT.createDP(profile);
This causes the program to hang. Any ideas why this may not work?
Thanks,
Henry Romijn
___
Hi,
I can't get the SearchProfile demo to work and I have tried various input
files. Is there a sample file available that will work so I can test if
something else is wrong?
Thanks,
Henry Romijn
Quoting [EMAIL PROTECTED]:
> Hi,
>
> I have tried to run the SearchProfile demo with something
Hi,
I have tried to run the SearchProfile demo with something simple like:
>a
MGNA
>b
MGNA
>c
MGNA
>d
MGNA
I use the ProfileHMM constructor that has a string as one of its parameters
(instead of the deprecated one). There appear to be no problems in creating
and initializing the ProfileHMM obj