Impact of sample size on the results

[alexximalayaunlv]

Hi there,

I'm wondering how the sample size would affect the accuracy of the results. Are there minimum sample size requirements? Let's say if I had four groups with 40, 30, 20, 10 sample size respectively, will the software perform well? Those are all metagenomics data. And I want to pick up the significant features (OTUs, genes) in those groups.

Thanks!!

[gailrosen]

Since this is not a wrapper method, it'll do the best it can, but yeah, I think it will be affected by class imbalance.

…

________________________________________ From: alexximalayaunlv <notifications@github.com> Sent: Monday, August 17, 2020 5:05 PM To: EESI/Fizzy Cc: Subscribed Subject: [EESI/Fizzy] Impact of sample size on the results (#6) External. Hi there, I'm wondering how the sample size would affect the accuracy of the results. Are there minimum sample size requirements? Let's say if I had four groups with 40, 30, 20, 10 sample size respectively, will the software perform well? Those are all metagenomics data. And I want to pick up the significant features (OTUs, genes) in those groups. Thanks!! — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub<https://nam01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FEESI%2FFizzy%2Fissues%2F6&data=02%7C01%7Cglr26%40drexel.edu%7C1f58902312e748e86ad808d842f14310%7C3664e6fa47bd45a696708c4f080f8ca6%7C0%7C0%7C637332951259787149&sdata=HFx8%2BO4K8zu57pGOPSYYRwmCVlqG86EFJzyqvWsdMRE%3D&reserved=0>, or unsubscribe<https://nam01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FABHZROLJQUOGWQLXSMDHM53SBGLRJANCNFSM4QCG23IQ&data=02%7C01%7Cglr26%40drexel.edu%7C1f58902312e748e86ad808d842f14310%7C3664e6fa47bd45a696708c4f080f8ca6%7C0%7C0%7C637332951259797145&sdata=BJnaYe19g6iopF%2BIB4yXsxRlLuNGpTo7YCtlSldQ8i8%3D&reserved=0>.

[alexximalayaunlv]

Thanks for the quick response! Do you think it's ok to filter the data using Fizzy and then use other non-parametric test to pick the deferentially abundant abundant genes among groups? I saw a research used mutual information to filter data and then use random forest to train the classifier. Thanks!!

[gailrosen]

you can try that i think

…

________________________________________ From: alexximalayaunlv <notifications@github.com> Sent: Monday, August 17, 2020 8:53 PM To: EESI/Fizzy Cc: Rosen,Gail; Comment Subject: Re: [EESI/Fizzy] Impact of sample size on the results (#6) External. Thanks for the quick response! Do you think it's necessary to filter the data using Fizzy and then use other non-parametric test to pick the deferentially abundant abundant genes among groups? I saw a research used mutual information to filter data and then use random forest to train the classifier. Thanks!! — You are receiving this because you commented. Reply to this email directly, view it on GitHub<https://nam01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FEESI%2FFizzy%2Fissues%2F6%23issuecomment-675187784&data=02%7C01%7Cglr26%40drexel.edu%7C36400f7d25f745ef3e4308d843111e85%7C3664e6fa47bd45a696708c4f080f8ca6%7C0%7C0%7C637333088092825736&sdata=MlRwc3xOowHoJ55G3EwJZ6g47SCIygtF8CHvVz%2BbAPA%3D&reserved=0>, or unsubscribe<https://nam01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FABHZROI5OI3MOCDYIMXVA2TSBHGINANCNFSM4QCG23IQ&data=02%7C01%7Cglr26%40drexel.edu%7C36400f7d25f745ef3e4308d843111e85%7C3664e6fa47bd45a696708c4f080f8ca6%7C0%7C0%7C637333088092835731&sdata=wuLONynnGmyPVL7hUHQJxMDLbLw63F6HfXE4MbpaMB4%3D&reserved=0>.