bash sort and paste columns in alphabetical order - bash

Question

I have a file.txt which has the following columns
id chr pos alleleA alleleB
1 01 1234 CT T
2 02 5678 G A
3 03 8901 T C
4 04 12345 C G
5 05 567890 T A
I am looking for a way of creating a new column so that it looks like : chr:pos:alleleA:alleleB
The problem is that alleleA and alleleB should be sorted based on:
1. alphabetical order
2. either of these two columns with more letter per line should be first and followed by the second column
In this example , it would look like this :
id chr pos alleleA alleleB newID
1 01 1234 CT T chr1:1234:CT:T
2 02 5678 G A chr2:5678:A:G
3 03 8901 T C chr3:8901:C:T
4 04 12345 C G chr4:12345:C:G
5 05 567890 T A chr5:567890:A:T
I appreciate any help and suggestion. Thanks.
EDIT
Up to now i can modify chr column so that it will have a look as "chr:1"...
AlleleA and AlleleB columns should be combined so that if either of columns contains more than 1 letter, in column newID it would be in the first place. If there is only one letter in both columns, these letters are arranged alphabetically in the newID column

Answer 1

gawk solution:
awk 'function custom_sort(i1,v1,i2,v2){ # custom function to compare 2 crucial fields
l1=length(v1); l2=length(v2); # getting length of both fields
if (l1 == l2) {
return (v1 > v2)? 1:-1 # compare characters if field lengths are equal
} else {
return l2 - l1 # otherwise - compare by length (descending)
}
} NR==1 { $0=$0 FS "newID" } # add new column
NR>1 { a[1]=$4; a[2]=$5; asort(a,b,"custom_sort"); # sort the last 2 columns using function `custom_sort`
$(NF+1) = sprintf("chr%s:%s:%s:%s",$1,$3,b[1],b[2])
}1' file.txt | column -t
The output:
id chr pos alleleA alleleB newID
1 01 1234 CT T chr1:1234:CT:T
2 02 5678 G A chr2:5678:A:G
3 03 8901 T C chr3:8901:C:T
4 04 12345 C G chr4:12345:C:G
5 05 567890 T A chr5:567890:A:T

Answer 2

Perl to the rescue:
perl -lane '
if (1 == $.) { print "$_ newID" }
else { print "$_ ", join ":", "chr" . ($F[1] =~ s/^0//r),
$F[2],
sort { length $b <=> length $a
or $a cmp $b
} #F[3,4];
}' -- input.txt
-l removes newlines from input and adds them to print
-n reads the input line by line
-a splits each input line on whitespace into the #F array
$. is the input line number, the condition just prints the header for the first line
s/^0// removes the initial zero from $F[1] (i.e. column 2)
/r returns the result of the substitution
the last two column lenghts are compared, if they are the same, string comparison is used.