CADASIL TabCAT

Notes

As you plan and work on your analyses, I wanted to provide a bit of context on the risk_score variable. In case you aren’t familiar with this specific pocket of the CADASIL literature, we stratify NOTCH3 variants into three pathogenic risk tiers based on the EGFr domain affected (following the aggregation method proposed by Hack et al. 

High Risk (risk_score = 9): EGFr domains 1–6, 8, 10, 11, 26

Medium Risk (risk_score = 8): EGFr domains 7, 9, 12–18

Low Risk (risk_score = 7): EGFr domains 19–25, 27–34

Control (risk_score = 0): NOTCH3 gene-negative family members

With that in mind, it would be fantastic to compare the age curves you’ve produced using these stratifications. Specifically, could we look at:

1. Case vs. Control (2 separate curves):

Control (Score = 0) vs. All CADASIL cases (Scores = 7, 8, or 9)

(Note: You may also want to run a quick comparison of just 0 vs. 9 to look at the most extreme contrast).

2. Full Risk-Level Stratification (4 separate curves):

Control (0) vs. Low (7) vs. Medium (8) vs. High (9)

In many of our other analyses so far, we are seeing a quantifiable, stepwise relationship (either Control < Low < Medium < High, or the exact inverse). It will be really interesting to see if the TabCAT data follows this same pattern!

Results split by High vs Control

Age X Risk Group Interaction (High, Control)

TabCAT EXAMINER ~ age + group + gender + education + age_squared*group

Term	Estimate	Std. Error	t value	Pr(>|t|)
(Intercept)	-0.3287545787	0.30479356979	-1.0786139	0.2813276135685614521
Education	0.0806957461	0.01096039135	7.3624877	0.0000000000008445974
Female(0) vs Male(1)	-0.1513624957	0.05333950185	-2.8377186	0.0047454570099816368
Age	0.0100401925	0.01183159763	0.8485914	0.3965519643019229656
Age Quadratic	-0.0002760713	0.00012898229	-2.1403816	0.0328514663497252377
Control(0) vs High(1) Risk	0.2003531242	0.12039381802	1.6641479	0.0967668435044460595
Control(0) vs High(1) Risk X Age Quadratic	-0.0001769512	0.00004274131	-4.1400517	0.0000413414465858735

Results split by Control (0) vs. Low (7) vs. Medium (8) vs. High (9)

Age X Risk Group Interaction (High, Medium, Low, Control)

TabCAT EXAMINER ~ age + group + gender + education + age_squared*group

Term	Estimate	Std. Error	t value	Pr(>|t|)
(Intercept)	-0.36929516890	0.30487059317	-1.2113178	0.2264037639835747295
Education	0.07990551218	0.01088511476	7.3408057	0.0000000000009887643
Female(0) vs Male(1)	-0.14502986430	0.05304232866	-2.7342288	0.0064971644307018428
Age	0.01240706172	0.01186630745	1.0455706	0.2963157086286456465
Age Quadratic	-0.00030074119	0.00012926754	-2.3265021	0.0204310341193957579
Control(0) vs Low(1) Risk	0.09056275967	0.25331172515	0.3575151	0.7208723830550470835
Control(0) vs Medium(1) Risk	0.13976718086	0.20024032981	0.6979972	0.4855363042424638564
Control(0) vs High(1) Risk	0.23448712481	0.12399247486	1.8911400	0.0592420128181554637
Control(0) vs Low(1) Risk X Age Quadratic	-0.00009409111	0.00009305967	-1.0110836	0.3125149792408505878
Control(0) vs Medium(1) Risk X Age Quadratic	-0.00010809851	0.00006401199	-1.6887228	0.0919592261224735086
Control(0) vs High(1) Risk X Age Quadratic	-0.00020833862	0.00004424880	-4.7083450	0.0000033227359897527