Estimate lambda

Questions about GenABEL (aka *ABEL) suite of packages
Forum rules
Please remember not to post any sensitive data on this public forum.
The first few posts of newly registered users will be moderated in order to filter out any spammers.

When get a solution to the problem you posted, please change the topic name (e.g. from "how to ..." to "[SOLVED] how to ..."). This will make it easier for the community to follow the posts yet to be attended.
hakosalo
GenABEL suite user
GenABEL suite user
Posts: 12
Joined: Thu May 12, 2011 7:45 am

Estimate lambda

Postby hakosalo » Fri Oct 14, 2011 8:08 am

I am running to this warning all the time with many of my analysis if I try to do association analysis
> Assoc1<-qtscore(pheno, own.data, trait="binomial")
Warning message:
In qtscore(pheno, own.data, trait = "binomial") : z2 lambda < 1, set to 1

In this case my own.data is not human data, but canine. Does that make a difference for this estimation? Does this matter anyway? I can run "polygenic" for this data set anyway, eventhough quite often if I have trait="binomial" it gives BAD CONVERGENCE which is kind of expected? If I have quantitative trait and BAD CONVERGENCE what should I do? (sample sizes are normally 20-60 cases with similar number of controls)
Thanks!
OH

MariaG
GenABEL specialist
GenABEL specialist
Posts: 36
Joined: Thu Feb 03, 2011 12:29 pm

Re: Estimate lambda

Postby MariaG » Wed Oct 19, 2011 2:52 pm

Here is the additional message from the same author (hakosalo):

hakosalo wrote:Hi.
I have this type of error message appearing over and over again (my data is not human, its canine)

"Testassoc<-qtscore(pheno, own.data, trait="binomial")
Warning message:
In qtscore(pheno, own.data, trait = "binomial") : z2 lambda < 1, set to 1"

Why is it in almost every data set I analyze give this message, and does this matter? Sample size is normally everything between 20-60 cases + same amount of controls.

Also, when I try to estimate the possible polygenic effect i get :
"
MyPol<-polygenic(pheno, own.data, kin=Bkin, trait="binomial")
LM estimates of fixed parameters:
[1] -0.09531018
iteration = 0
Step:
[1] 0 0 0

"

and so on all the way to the message :

difFGLS:
[1] 0.495997
fixed effect betas changed to FGLS-betas for re-estimation

***********************************************
*** !!!BAD!!! convergence indicated by FGLS ***
***********************************************

Now, is this because I have binary trait or what?


Anyway I try to test also mixed model , I get the same lamda message:


> MixTest<-qtscore(MyPol$pgres, own.data)
Warning message:
In qtscore(MyPol$pgres, belgit) : z2 lambda < 1, set to 1

Why doesn't the program accept lambda which is smaller than one? Why am I getting it all the time? The data should be ok. Does it make a difference that I don't have human, but canine data?

Thanks for your answers!

Osmo

yurii
GenABEL developer
GenABEL developer
Posts: 263
Joined: Fri Jan 21, 2011 5:20 pm

Re: Estimate lambda

Postby yurii » Sat Oct 29, 2011 12:18 pm

Re: z1<1, see

http://forum.genabel.org/viewtopic.php?f=6&t=484

With polygenic, do not use "binomial" option -- see

http://forum.genabel.org/viewtopic.php?f=6&t=149

and

http://forum.genabel.org/viewtopic.php?f=6&t=537

for more details

(and there may be few more posts on that topic...)
Note that (Gen)ABELs are dynamically developing; while this post is intended to provide full information at the time of posting, please read on further posts, if any, as the topic may be updated with novel solutions at a later stage.

best regards,
Yurii


Return to “GenABEL”

Who is online

Users browsing this forum: No registered users and 1 guest