Question: Minia was killed when counting kmer and abundance-min choice
gravatar for caizexi123
17 months ago by
caizexi12340 wrote:

Hi guys,

I am running minia on two machine. One of them was killed during counting kmer. The machine is ~700GB memory and I have ITB disk for the output. The genome is human-size with 100x data. The data was trimed and error corrected. The last message of the log is "MB EXCEPTION: Pool allocation failed for 74 bytes (bank ids alloc). Current usage is 4154491796 and capacity is 4154491854". What is the problem? I tried three times, it is the same. I even recompile the minia.

The second one is a plant genome about 1,7GB, also with 100x. The data was trimed and error corrected. I used the kmergenie to find the best k, is 91 and the coverage cut-off is 1. But I run minia with default abundance-min, and the result is very bad. Should I set abundance-min to 1? And why the kmergenie suggest coverage cut-off is 1? Is it because all the preprocess I have done?


minia • 711 views
ADD COMMENTlink written 17 months ago by caizexi12340

Which version of Minia? Which version of KmerGenie?

See here, you probably want to increase min_abundance.

ADD REPLYlink written 17 months ago by h.mon4.7k

Given the error "Pool alloc.." reported, it is a newer version than in then previous link.

- How many input files did you give, is it a large number?

- Coverage cutoff of 1 is always going to give bad results, I recommend at least 2. Kmergenie may recommend 1 if it sees that the coverage is low. What do the kmergenie histograms look like?

ADD REPLYlink modified 16 months ago • written 16 months ago by Rayan Chikhi1.1k

Sorry, I haven't see the reply (new in biostar).

1、I am sure the number of input files is below 10.

2、But the sequencing coverage is more than 100. Is it caused by I keep all the short reads after trimming? Should I remove all the reads shorter than the recommend kmer?

ADD REPLYlink written 16 months ago by caizexi12340

Hi, sorry. I don't know my question have answer.

Minia is 2.0.3, Kmergenie is 1.6982.

I tried the default (2) and the one kmergenis suggested (1), both give me very bad results. The total length just cover half of the genome.


ADD REPLYlink written 16 months ago by caizexi12340
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 515 users visited in the last hour