# MMUPHin lm\_meta: Questions regarding warning messages and forest plot generate

**URL:** <https://forum.biobakery.org/t/mmuphin-lm-meta-questions-regarding-warning-messages-and-forest-plot-generate/6596>\
**Category:** MMUPHin\
**Created:** [February 6, 2024, 1:50pm UTC](https://forum.biobakery.org/t/mmuphin-lm-meta-questions-regarding-warning-messages-and-forest-plot-generate/6596 "2024-02-06T13:50:28Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![iordanis\_theocharis](https://yyz1.discourse-cdn.com/flex027/user_avatar/forum.biobakery.org/iordanis_theocharis/32/2709_2.png) [@iordanis\_theocharis](https://forum.biobakery.org/u/iordanis_theocharis)\
**Post date:** [February 6, 2024, 1:50pm UTC](https://forum.biobakery.org/t/mmuphin-lm-meta-questions-regarding-warning-messages-and-forest-plot-generate/6596/1 "2024-02-06T13:50:28Z")

</div>

Hello,

I am new born in field of bioinformatics and i am glad to using MMUPHin R packages for meta-analysis of gut microbiome 16S sequencing data.

I use this tool to detect different abundance microbes based in different disease states, compare with healthy controls.

I have 3 questions regardless MMPHin tool.

1. i run the lm\_meta after batch/study correct,

data\_meta$Host\_case\_disease ← relevel(x = as.factor(data\_meta$Host\_case\_disease), ref = “disease\_type”)

I have not another variables to check only the different type of disease

fit\_lm\_meta ← lm\_meta(feature\_abd = data\_abd\_adj,  
batch = “Bioproject”,  
exposure = “Virus”,  
covariates = c(“Host\_case\_disease”),  
data = data\_meta,  
control = list(verbose = FALSE))

meta\_fits ← fit\_lm\_meta$meta\_fits

my question is, if someone knows what is mean this warnings ?

Fisher scoring algorithm may have gotten stuck at a local maximum.  
Setting tau^2 = 0. Check the profile likelihood plot with profile().

I cannot find something that explain how to explain me for Fisher algorithm or i am not so good to find something. The only think i know is the tau^=0 mean no heterogeneity so i can combine the datasets.

1. I use the same code with out covariates and the warning just disappear  
so my question is, can i don’t use covariates ?

2. can i produce forest plot based random effect model based in MMPHin tool?  
so i can compare the different state of disease like paired-test

I am sorry for so match question, If anyone can help, it would be greatly appreciated!!!

Thanks for your support