06 - Cutaneous Neurofibroma patient Variables
Sara Gosline
2019-10-08
Synapse tables with data in them
We have two tables on synapse that have cNF patient data in them - Patient Table - Sample Table
pat.tab<-synapser::synTableQuery("SELECT Patient,Race,Gender,Age,Pain,Itching FROM syn7342635")$asDataFrame()%>%
select(-c(ROW_ID,ROW_VERSION))
samp.tab<-synapser::synTableQuery("SELECT Patient,TumorNumber,Length_in_mm FROM syn5556216")$asDataFrame()%>%
select(-c(ROW_ID,ROW_VERSION))
pat.tab## Patient Race Gender Age Pain Itching
## 1 1 White Female 48 FALSE TRUE
## 2 2 Asian Female 34 FALSE FALSE
## 3 3 Hispanic Female 48 FALSE FALSE
## 4 4 White Female 36 FALSE TRUE
## 5 5 White Male 65 FALSE FALSE
## 6 6 White Female 46 FALSE TRUE
## 7 8 Black Female 27 FALSE FALSE
## 8 9 White Male 35 FALSE FALSE
## 9 10 White Male 58 FALSE FALSE
## 10 11 Black Female 30 TRUE TRUE
## 11 13 White Female 41 TRUE TRUEsamp.tab## Patient TumorNumber Length_in_mm
## 1 1 3 NA
## 2 1 4 14
## 3 1 6 14
## 4 1 9 15
## 5 1 1 15
## 6 1 2 13
## 7 2 4 7
## 8 2 6 6
## 9 2 7 5
## 10 2 10 7
## 11 2 1 5
## 12 2 2 7
## 13 2 8 9
## 14 2 9 7
## 15 3 1 11
## 16 3 2 16
## 17 3 3 20
## 18 3 4 15
## 19 4 1 8
## 20 4 10 7
## 21 4 4 5
## 22 4 9 10
## 23 5 11 10
## 24 5 12 9
## 25 5 13 11
## 26 5 15 9
## 27 5 5 8
## 28 5 8 6
## 29 6 4 15
## 30 6 5 10
## 31 6 6 15
## 32 6 7 13
## 33 8 1 10
## 34 8 5 13
## 35 8 7 12
## 36 8 9 5
## 37 8 4 10
## 38 8 6 10
## 39 9 1 12
## 40 9 6 5
## 41 9 7 5
## 42 9 10 8
## 43 10 1 11
## 44 10 2 20
## 45 10 3 10
## 46 10 5 20
## 47 10 4 5
## 48 11 1 13
## 49 11 7 10
## 50 11 8 10
## 51 11 14 20
## 52 11 2 5
## 53 11 3 3
## 54 13 1 18
## 55 13 2 7
## 56 13 3 12
## 57 13 7 6
## 58 1 NA NA
## 59 2 NA NA
## 60 3 NA NA
## 61 4 NA NA
## 62 5 NA NA
## 63 6 NA NA
## 64 8 NA NA
## 65 9 NA NA
## 66 10 NA NA
## 67 11 NA NA
## 68 9 3 13For the purposes of this analysis we want to have just the age, sex, tumor size and and ‘reformed’ patient name
full.tab<-samp.tab%>%left_join(pat.tab,by='Patient')%>%
mutate(specimenID=paste0('patient',Patient,'tumor',TumorNumber))
full.tab## Patient TumorNumber Length_in_mm Race Gender Age Pain Itching
## 1 1 3 NA White Female 48 FALSE TRUE
## 2 1 4 14 White Female 48 FALSE TRUE
## 3 1 6 14 White Female 48 FALSE TRUE
## 4 1 9 15 White Female 48 FALSE TRUE
## 5 1 1 15 White Female 48 FALSE TRUE
## 6 1 2 13 White Female 48 FALSE TRUE
## 7 2 4 7 Asian Female 34 FALSE FALSE
## 8 2 6 6 Asian Female 34 FALSE FALSE
## 9 2 7 5 Asian Female 34 FALSE FALSE
## 10 2 10 7 Asian Female 34 FALSE FALSE
## 11 2 1 5 Asian Female 34 FALSE FALSE
## 12 2 2 7 Asian Female 34 FALSE FALSE
## 13 2 8 9 Asian Female 34 FALSE FALSE
## 14 2 9 7 Asian Female 34 FALSE FALSE
## 15 3 1 11 Hispanic Female 48 FALSE FALSE
## 16 3 2 16 Hispanic Female 48 FALSE FALSE
## 17 3 3 20 Hispanic Female 48 FALSE FALSE
## 18 3 4 15 Hispanic Female 48 FALSE FALSE
## 19 4 1 8 White Female 36 FALSE TRUE
## 20 4 10 7 White Female 36 FALSE TRUE
## 21 4 4 5 White Female 36 FALSE TRUE
## 22 4 9 10 White Female 36 FALSE TRUE
## 23 5 11 10 White Male 65 FALSE FALSE
## 24 5 12 9 White Male 65 FALSE FALSE
## 25 5 13 11 White Male 65 FALSE FALSE
## 26 5 15 9 White Male 65 FALSE FALSE
## 27 5 5 8 White Male 65 FALSE FALSE
## 28 5 8 6 White Male 65 FALSE FALSE
## 29 6 4 15 White Female 46 FALSE TRUE
## 30 6 5 10 White Female 46 FALSE TRUE
## 31 6 6 15 White Female 46 FALSE TRUE
## 32 6 7 13 White Female 46 FALSE TRUE
## 33 8 1 10 Black Female 27 FALSE FALSE
## 34 8 5 13 Black Female 27 FALSE FALSE
## 35 8 7 12 Black Female 27 FALSE FALSE
## 36 8 9 5 Black Female 27 FALSE FALSE
## 37 8 4 10 Black Female 27 FALSE FALSE
## 38 8 6 10 Black Female 27 FALSE FALSE
## 39 9 1 12 White Male 35 FALSE FALSE
## 40 9 6 5 White Male 35 FALSE FALSE
## 41 9 7 5 White Male 35 FALSE FALSE
## 42 9 10 8 White Male 35 FALSE FALSE
## 43 10 1 11 White Male 58 FALSE FALSE
## 44 10 2 20 White Male 58 FALSE FALSE
## 45 10 3 10 White Male 58 FALSE FALSE
## 46 10 5 20 White Male 58 FALSE FALSE
## 47 10 4 5 White Male 58 FALSE FALSE
## 48 11 1 13 Black Female 30 TRUE TRUE
## 49 11 7 10 Black Female 30 TRUE TRUE
## 50 11 8 10 Black Female 30 TRUE TRUE
## 51 11 14 20 Black Female 30 TRUE TRUE
## 52 11 2 5 Black Female 30 TRUE TRUE
## 53 11 3 3 Black Female 30 TRUE TRUE
## 54 13 1 18 White Female 41 TRUE TRUE
## 55 13 2 7 White Female 41 TRUE TRUE
## 56 13 3 12 White Female 41 TRUE TRUE
## 57 13 7 6 White Female 41 TRUE TRUE
## 58 1 NA NA White Female 48 FALSE TRUE
## 59 2 NA NA Asian Female 34 FALSE FALSE
## 60 3 NA NA Hispanic Female 48 FALSE FALSE
## 61 4 NA NA White Female 36 FALSE TRUE
## 62 5 NA NA White Male 65 FALSE FALSE
## 63 6 NA NA White Female 46 FALSE TRUE
## 64 8 NA NA Black Female 27 FALSE FALSE
## 65 9 NA NA White Male 35 FALSE FALSE
## 66 10 NA NA White Male 58 FALSE FALSE
## 67 11 NA NA Black Female 30 TRUE TRUE
## 68 9 3 13 White Male 35 FALSE FALSE
## specimenID
## 1 patient1tumor3
## 2 patient1tumor4
## 3 patient1tumor6
## 4 patient1tumor9
## 5 patient1tumor1
## 6 patient1tumor2
## 7 patient2tumor4
## 8 patient2tumor6
## 9 patient2tumor7
## 10 patient2tumor10
## 11 patient2tumor1
## 12 patient2tumor2
## 13 patient2tumor8
## 14 patient2tumor9
## 15 patient3tumor1
## 16 patient3tumor2
## 17 patient3tumor3
## 18 patient3tumor4
## 19 patient4tumor1
## 20 patient4tumor10
## 21 patient4tumor4
## 22 patient4tumor9
## 23 patient5tumor11
## 24 patient5tumor12
## 25 patient5tumor13
## 26 patient5tumor15
## 27 patient5tumor5
## 28 patient5tumor8
## 29 patient6tumor4
## 30 patient6tumor5
## 31 patient6tumor6
## 32 patient6tumor7
## 33 patient8tumor1
## 34 patient8tumor5
## 35 patient8tumor7
## 36 patient8tumor9
## 37 patient8tumor4
## 38 patient8tumor6
## 39 patient9tumor1
## 40 patient9tumor6
## 41 patient9tumor7
## 42 patient9tumor10
## 43 patient10tumor1
## 44 patient10tumor2
## 45 patient10tumor3
## 46 patient10tumor5
## 47 patient10tumor4
## 48 patient11tumor1
## 49 patient11tumor7
## 50 patient11tumor8
## 51 patient11tumor14
## 52 patient11tumor2
## 53 patient11tumor3
## 54 patient13tumor1
## 55 patient13tumor2
## 56 patient13tumor3
## 57 patient13tumor7
## 58 patient1tumorNA
## 59 patient2tumorNA
## 60 patient3tumorNA
## 61 patient4tumorNA
## 62 patient5tumorNA
## 63 patient6tumorNA
## 64 patient8tumorNA
## 65 patient9tumorNA
## 66 patient10tumorNA
## 67 patient11tumorNA
## 68 patient9tumor3Now we can evaluate other things
Plot the data
##now what do we see on a tissue level?
require(ggplot2)
p<-ggplot(full.tab)+geom_point(aes(x=Age,y=Length_in_mm,color=Itching,shape=Gender))+ggtitle("Age by tumor size with itching")
print(p)## Warning: Removed 11 rows containing missing values (geom_point).
p<-ggplot(full.tab)+geom_point(aes(x=Age,y=Length_in_mm,color=Itching,shape=Pain))+ggtitle("Age by tumor size with pain")
print(p)## Warning: Removed 11 rows containing missing values (geom_point).
So the pain/itching variables only go by patient, so we can’t figure out which samples are in pain/itching.
Get the expression data and see what correlates
exp.data<-synTableQuery('select * from syn20449214 where tumorType=\'Cutaneous Neurofibroma\'')$asDataFrame()
data.with.var<-exp.data%>%left_join(full.tab,by='specimenID')
#now compute the correlation with size for each transcript...?
gene.cors=data.with.var%>%group_by(Symbol)%>%mutate(corVal=cor(zScore,Length_in_mm))
top.genes=select(gene.cors,corVal)%>%distinct()%>%arrange(desc(corVal))%>%select(Symbol)## Adding missing grouping variables: `Symbol`bottom.genes=select(gene.cors,corVal)%>%distinct()%>%arrange(corVal)%>%select(Symbol)## Adding missing grouping variables: `Symbol`Now we have the genes most correlated with tumor size
top.genes## # A tibble: 19,098 x 1
## # Groups: Symbol [19,098]
## Symbol
## <chr>
## 1 AKAP12
## 2 SOCS2
## 3 DAPK2
## 4 SVEP1
## 5 RGL1
## 6 DTX4
## 7 HOXB3
## 8 ADAMTS5
## 9 VPS35L
## 10 ABI3BP
## # … with 19,088 more rowsand the genes least correlated
bottom.genes## # A tibble: 19,098 x 1
## # Groups: Symbol [19,098]
## Symbol
## <chr>
## 1 DAPL1
## 2 BPNT1
## 3 GSKIP
## 4 RPUSD3
## 5 ZNF707
## 6 IGHMBP2
## 7 WDR12
## 8 IPO13
## 9 MARCH9
## 10 SP6
## # … with 19,088 more rowsNow what do these look like?
ggplot(subset(data.with.var,Symbol%in%c(top.genes$Symbol[1:10])))+geom_point(aes(x=Length_in_mm,y=zScore,col=Symbol,shape=sex))+ggtitle('10 most correlated genes')
ggplot(subset(data.with.var,Symbol%in%c(bottom.genes$Symbol[1:10])))+geom_point(aes(x=Length_in_mm,y=zScore,col=Symbol,shape=sex))+ggtitle('10 most anti-correlated genes')
We should do pathway enrichment of these genes, see if they are doing anything. The fact that HOXB3 is there is already interesting. What else could there be???
Next steps
I think this is actually pretty cool - we can try to identify which are the ‘driving’ genes of cNF growth. * What pathways are enriched in these genes? * Are these correlations statistically significant? * Are there differences in correlated genes b/w male and female patients? * What are these genes doing in pNFs and MPNSTs? * Are any of these correlated with immune reponse (conversely are immune signatures correlated with this size variable)?
Are there other questions we can answer?