Hi,
As suggested by many others, SCALE performs better than KR in most cases, so I want to use addNorm to add the SCALE normalizations. When I check the usage using the command "java -jar juicer_tools_1.22.01.jar addNorm", it shows:
Usage: juicer_tools addNorm <input_HiC_file> [input_vector_file]
: -d use intra chromosome (diagonal) [false]
: -F don't calculate normalization for fragment-delimited maps [false]
: -w <int> calculate genome-wide resolution on all resolutions >= input resolution [not set]
Above options ignored if input_vector_file present
: -k normalizations to include
: -r resolutions for respective normalizations to build to
: -j number of CPU threads to use
Then I tried the command "java -Xmx300G -jar juicer_tools_1.22.01 addNorm -d -F -k KR,VC_SQRT,SCALE -w 10000 -r 10000,10000,10000 -j 40 $hicfile"
and after dump I could get 10kb contacts with SCALE, VC_SQRT or KR, but not other resolutions such as 50kb with any normalization methods.
My question is how could I get the contacts at multiple resolutions?
I tried the command "java -Xmx300G -jar juicer_tools_1.22.01 addNorm -d -F -k SCALE,SCALE -w 50000 -r 100000,50000 -j 40 $hicfile"
and only get 50kb but not 100kb contacts after dump.
I also tried the command "java -Xmx300G -jar juicer_tools_1.22.01 addNorm -d -F -k SCALE,SCALE -w 50000 -r 50000,100000 -j 40 $hicfile"
and only get 100kb but not 50kb contacts after dump.
Is it because the addNorm only considers the last resolution for the normalization method? Then can I get multiple resolutions for the same normalization method? And how shall I know which resolution(s) I have in the .hic files?
Many thanks for any help in advance!
Kelly