The same reads were not polished

[cym0304]

Hi, I created a test file like this:

@test1
ATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCG
+
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

The length of this read is 100 and I copied it 5 times and changed the IDs. Finally, the test file has 5 reads which has the same sequence and different name. Than I test the racon with the command like this:

minimap2 ./test2.fq ./test2.fq -a -t 4 > test2.align.sam racon ./test2.fq ./test2.align.sam ./test2.fq > test2_consensus.fa

But the consensus reads like this:

>test2 LN:i:100 RC:i:2 XC:f:1.000000
ATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCG
>test4 LN:i:100 RC:i:2 XC:f:1.000000
ATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCGATCG

In my imagination, there will be only one sequences in the result file. But I got two same consensus reads and the RC value shows that only 4 raw reads were polished. Then I set the -w as 10 or 50, it did not work.

What the reason of this situation? And could you tell me what should I do to get the correct result? Thank you.

[rvaser]

Hello, what is the purpose of this test you conducted? Do you want to test read error correction or contig polishing?

Best regards, Robert

[cym0304]

Hello， I want to test the contig polishing. I think the repeat of a same sequence should be clustered to one consensus.

Best regards, cym

[rvaser]

This might occur but there is no guarantee for it. If you have repetitive regions without any overhangs left or right, it is highly likely that there will be multiple consensus sequences with that region because racon picks the best (longest) overlap for each read and then polishes each target sequence. No other procedures are invoked.

Best regards, Robert