error when trying brnb for negative binomial

[cperk]

Hello,

I was excited to find this package to solve my complete separation problem. My count data is overdispersed so need to use the negative binomial family, which I understand is run using brnb(). I keep getting this error: NA/NaN/Inf in foreign function call (arg 1), even though I am specifying all the necessary terms:

salmonella_fm <- production ~ Treat_comboSource_Salinity +Rarity+ Region+Treat_comboTurion_length_num+replanted.x

N_T_6_2020_nbn <- brglm2::brnb(salmonella_fm,
data=turion_harvest_2021_all_gen_initial_turion,link = "log",transformation = "inverse", type = "AS_mean")

I've noticed that the error disappears if I use type="ML" but all other types trigger the error.

Could you please advise?

Thanks!

-Carrie

[ikosmidis]

Hi @cperk: thanks for the report. If you share a reproducible example (with data and code) that demonstrates the issue I can investigate. The reprex R package may come in handy. -I

[cperk]

Thanks!

Here is a reproducible example:

df1 <- data.frame(production=c(15,12,10,9,6,8,9,5,3,3,2,1,0,0,0,0), Treatment_Num=c(1,1,1,1,2,2,2,2,3,3,3,3,4,4,4,4), Treatment=c("Fresh","Fresh","Fresh","Fresh","low","low","low","low","med","med","med","med","high","high","high","high"),Genotype=c(1,1,2,2,1,1,2,2,1,1,2,2,1,1,2,2), Source_Salinity=c("Fresh","Fresh","Brackish","Brackish","Fresh","Fresh","Brackish","Brackish","Fresh","Fresh","Brackish","Brackish","Fresh","Fresh","Brackish","Brackish"), Days_to_death <- c(500, 500, 500, 500, 400, 350, 300, 500, 200, 202, 260, 280, 150,150,160,140), censored <- c(1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0))


salmonella_fm <- production ~ Treatment
N_T_6_2020_nbn_toy <- brglm2::brnb(salmonella_fm,
                               data=df1,link = "log",transformation = "inverse", type = "AS_mean")

I figured out that it only happens when the fixed effect is categorical; when I use a numeric version of Treatment, the error disappears:

salmonella_fm <- production ~ Treatment_Num
N_T_6_2020_nbn_toy <- brglm2::brnb(salmonella_fm,
                               data=df1,link = "log",transformation = "inverse", type = "AS_mean")

In my real data it wouldn't be appropriate to convert Treatment to numeric, so I would really appreciate it if you could find a solution for using a categorical fixed effect!

[cperk]

Any update on this?

[ikosmidis]

We are investigating the issue you opened. I understand that you may need a resolution quickly, but please note that we cannot guarantee the timeframe of response/resolution.

[cperk]

Hi! I totally understand that you couldn't resolve this issue right away; are you any closer?

[cperk]

any update?

[melwarden77]

I am having the same issue.