I have performed MANCOVAN toolbox to find the difference between two groups.
When I tried to display Mancovan results (univariate results), I got some points which I want to make clear.
1) As 'Threshold criteria (univariate results) : FDR' and ' P-threshold (univariate results): 0.01' , I got Figure 'Significant effects of GroupA - GroupB (p<0.01)' with the correlation diagram among components. Does it mean that these components are showing the group difference significantly at FDR corrected p<0.01, or FDR corrected (p<0.05) with uncorrected p<0.01 ?
2) If I set up as 'Threshold criteria (univariate results) : None' and ' P-threshold (univariate results): 0.01', does it mean that significantly at uncorrected p<0.01?
Thanks,
JC
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I have performed MANCOVAN toolbox to find the difference between two groups.
When I tried to display Mancovan results (univariate results), I got some points which I want to make clear.
1) As 'Threshold criteria (univariate results) : FDR' and ' P-threshold (univariate results): 0.01' , I got Figure 'Significant effects of GroupA - GroupB (p<0.01)' with the correlation diagram among components. Does it mean that these components are showing the group difference significantly at FDR corrected p<0.01, or FDR corrected (p<0.05) with uncorrected p<0.01 ?
2) If I set up as 'Threshold criteria (univariate results) : None' and ' P-threshold (univariate results): 0.01', does it mean that significantly at uncorrected p<0.01?
CAUTION: This email was sent from someone outside of the university. Do not click links or open attachments unless you recognize the sender and know the content is safe.
I have another question about FDR correction.
After validating components, I have decided to use 12 network, and I have performed MANCOVAN toolbox to find the difference between two groups using network. When I tried to display Mancovan results (univariate results) with FDR correction option, does this FDR correct over the number of components I am using (in my case, this is 12 networks)?
If so, if I change the number of components for this MANCOVAN, the significant results with FDR correction can be changed?
Thanks,
JC
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi GIFT expert!
I have performed MANCOVAN toolbox to find the difference between two groups.
When I tried to display Mancovan results (univariate results), I got some points which I want to make clear.
1) As 'Threshold criteria (univariate results) : FDR' and ' P-threshold (univariate results): 0.01' , I got Figure 'Significant effects of GroupA - GroupB (p<0.01)' with the correlation diagram among components. Does it mean that these components are showing the group difference significantly at FDR corrected p<0.01, or FDR corrected (p<0.05) with uncorrected p<0.01 ?
2) If I set up as 'Threshold criteria (univariate results) : None' and ' P-threshold (univariate results): 0.01', does it mean that significantly at uncorrected p<0.01?
Thanks,
JC
Hi JUngwon,
There is an option to select threshold criteria (fdr or none) in display mancova gui.
Thanks,
Srinivas
From: Jungwon Cha jungcha@users.sourceforge.net
Sent: Friday, May 6, 2022 8:21 AM
To: [icatb:discussion] 309910@discussion.icatb.p.re.sourceforge.net
Subject: [icatb:discussion] FDR correction in MANCOVAN toolbox
Hi GIFT expert!
I have performed MANCOVAN toolbox to find the difference between two groups.
When I tried to display Mancovan results (univariate results), I got some points which I want to make clear.
1) As 'Threshold criteria (univariate results) : FDR' and ' P-threshold (univariate results): 0.01' , I got Figure 'Significant effects of GroupA - GroupB (p<0.01)' with the correlation diagram among components. Does it mean that these components are showing the group difference significantly at FDR corrected p<0.01, or FDR corrected (p<0.05) with uncorrected p<0.01 ?
2) If I set up as 'Threshold criteria (univariate results) : None' and ' P-threshold (univariate results): 0.01', does it mean that significantly at uncorrected p<0.01?
Thanks,
JC
FDR correction in MANCOVAN toolboxhttps://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsourceforge.net%2Fp%2Ficatb%2Fdiscussion%2F309910%2Fthread%2F93525d9636%2F%3Flimit%3D25%23b8b8&data=05%7C01%7Csrachakonda%40gsu.edu%7C195fc145571f443f914208da2f6c5563%7C515ad73d8d5e4169895c9789dc742a70%7C0%7C0%7C637874440321336822%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=xrmZ%2Bt6EkypAv5YUUM1k8sjbkfEsEFDFQMlcDn6TwnQ%3D&reserved=0
Sent from sourceforge.net because you indicated interest in https://sourceforge.net/p/icatb/discussion/309910/https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsourceforge.net%2Fp%2Ficatb%2Fdiscussion%2F309910%2F&data=05%7C01%7Csrachakonda%40gsu.edu%7C195fc145571f443f914208da2f6c5563%7C515ad73d8d5e4169895c9789dc742a70%7C0%7C0%7C637874440321336822%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=6FsJBC3sKjPxyGMK%2FAm5dbE5Qr7R%2BD6y78kGjl6%2FaVw%3D&reserved=0
To unsubscribe from further messages, please visit https://sourceforge.net/auth/subscriptions/https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsourceforge.net%2Fauth%2Fsubscriptions%2F&data=05%7C01%7Csrachakonda%40gsu.edu%7C195fc145571f443f914208da2f6c5563%7C515ad73d8d5e4169895c9789dc742a70%7C0%7C0%7C637874440321336822%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=nEKpH%2FbSe0DQ20kQX2nkMK3emwl%2Fp5WMnGEW%2FwBDdJU%3D&reserved=0
CAUTION: This email was sent from someone outside of the university. Do not click links or open attachments unless you recognize the sender and know the content is safe.
Thank you!!
JC
Hi Srinivas,
I have another question about FDR correction.
After validating components, I have decided to use 12 network, and I have performed MANCOVAN toolbox to find the difference between two groups using network. When I tried to display Mancovan results (univariate results) with FDR correction option, does this FDR correct over the number of components I am using (in my case, this is 12 networks)?
If so, if I change the number of components for this MANCOVAN, the significant results with FDR correction can be changed?
Thanks,
JC