could Hmisc/rcorr change to replace NA to 0 if there are just 1 NA in the pair?

[BowenXiao123]

hi,

could Hmisc/rcorr change to replace NA to 0 if there are just 1 NA in the pair and drop the pair which contains two NA?

as we know, Hmisc/rcorr function can calculate the correlation between unNaN pairs. but here is a potential problem that is I have a gene expression quantification level list and a miRNA expression quantification level list

• NA 5.32 NA 6.45 NA 7.23
• 3.34 NA 5.43 NA 4.15 NA

when I am trying to calculate the expression correlation between the gene and miRNA it will return NA. but seems like there are correlated.

you may say why I do not change the NA to 0. because I have to avoid another situation that

• NA NA 5.22 NA NA NA
• NA NA 5.43 NA NA NA

if I replace NA with 0, both 0 pairs will affect the correlation coefficient result, because there is just one actually points in the data.

thank you for your attention. Have a nice day.

[BowenXiao123]

and there is another situation that is if i have

• 3.56 3.78 3.45 3.65 3.92
• NA 2.34 NA 2.65 NA NA

the rcorr will return a high correlation between them because it dropped 1 NA pairs and just remain pairs(3.78 2.34) and (3.65 2.65). but actually, it seems like there is no correlation between them if we replace NA with 0.