ValueError: The first argument of bincount must be non-negative

[zhanyinx]

Hi Fidel,

I am trying to use HiCAssemble to build a genome from Hi-C data. I have analysed the data using HiC-Pro and converted the output to h5 format using HiCExplorer 3.0

Then I used HiCAssemble to build the new genome as follow:

assemble -m goat_2019_july_120000_abs.h5 -o assembled_genome_goat --min_scaffold_length 100000 --bin_size 120000 --misassembly_zscore_threshold -1.0 --num_iterations 3 --num_processors 10 -f goat_v1_scaffolds.fa

This gave me the following error:

/work2/ggiorget/zhan/bin/Python-2.7.12/lib/python2.7/site-packages/hicexplorer/HiCMatrix.py:819: RuntimeWarning: divide by zero encountered in true_divide
dist_list = (np.array(dist_list).astype(float) / binsize).astype(int) + 1
/work2/ggiorget/zhan/bin/Python-2.7.12/lib/python2.7/site-packages/hicexplorer/HiCMatrix.py:819: RuntimeWarning: invalid value encountered in true_divide
dist_list = (np.array(dist_list).astype(float) / binsize).astype(int) + 1
Traceback (most recent call last):
File "/work2/ggiorget/zhan/bin/Python-2.7.12/bin/assemble", line 312, in
main(args)
File "/work2/ggiorget/zhan/bin/Python-2.7.12/bin/assemble", line 304, in main
scaffolds_to_ignore=args.scaffolds_to_ignore)
File "/work2/ggiorget/zhan/bin/Python-2.7.12/lib/python2.7/site-packages/hicassembler/HiCAssembler.py", line 88, in init
self.load_hic_matrix(hic_file_name, split_misassemblies, split_positions_file, matrix_bin_size)
File "/work2/ggiorget/zhan/bin/Python-2.7.12/lib/python2.7/site-packages/hicassembler/HiCAssembler.py", line 141, in load_hic_matrix
self.split_misassemblies(hic_file_name, split_positions_file)
File "/work2/ggiorget/zhan/bin/Python-2.7.12/lib/python2.7/site-packages/hicassembler/HiCAssembler.py", line 848, in split_misassemblies
ft.hic_ma.matrix = ft.hic_ma.convert_to_obs_exp_matrix(perchr=True)
File "/work2/ggiorget/zhan/bin/Python-2.7.12/lib/python2.7/site-packages/hicexplorer/HiCMatrix.py", line 822, in convert_to_obs_exp_matrix
sum_counts = np.bincount(dist_list, weights=submatrix.data)
ValueError: The first argument of bincount must be non-negative

HiCExplorer version on server: 2.1.1
HiCAssembler on server: last version
python version 2.7.12

Do you have any idea why this happend?

Thanks for your help
best
Zhan

[fidelram]

Converting the data from HiC pro to HiC explorer may not work properly because of the structure of the data. Can you try producing the data directly with HiCExplorer?