chhylp123 / hifiasm

Hifiasm: a haplotype-resolved assembler for accurate Hifi reads
MIT License
547 stars 87 forks source link

Error in void recalcate_window_advance #559

Closed kokyriakidis closed 1 year ago

kokyriakidis commented 1 year ago

Hi @chhylp123!

I got the following error:

[M::ha_hist_line]  rest: ***** 436494
[M::ha_analyze_count] left: none
[M::ha_analyze_count] right: none
[M::ha_pt_gen] peak_hom: 8; peak_het: -1
[M::ha_ct_shrink::603.467*15.42] ==> counted 83493907 distinct minimizer k-mers
[M::ha_pt_gen::] counting in normal mode
[M::yak_count] collected 752265965 minimizers
[M::ha_pt_gen::681.746*18.11] ==> indexed 748094456 positions, counted 83493907 distinct minimizer k-mers
hifiasm: Correct.cpp:3292: void recalcate_window_advance(overlap_region_alloc*, All_reads*, const ul_idx_t*, UC_Read*, Correct_dumy*, UC_Read*, kvec_t_u64_warp*, int64_t, double, double): Assertion `error != (unsigned int)-1' failed.
/var/lib/slurm/slurmd/job1419495/slurm_script: line 67: 2781789 Aborted                 (core dumped) hifiasm -o "${OUTPUT_PATH}/1_HIFIASM_ASSEMBLIES/${SAMPLE_NAME}/${SAMPLE_NAME}.asm" -t "${THREADS}" --ul "${HIFIASM_UL_FASTQ},${HIFIASM_SIMPLEX_FASTQ}" --ul-rate 0.1 --ul-cut 50000 "${HIFIASM_DUPLEX_FASTQ}"

I used EC reads from another tool. Do you think this error happens because of the new quality values?

chhylp123 commented 12 months ago

@kokyriakidis Which version are you using? The latest version should already fix it.

kokyriakidis commented 12 months ago

@chhylp123 I removed IUPAC characters and made it work. Using the latest version, can I make it work with datasets with IUPAC characters?

chhylp123 commented 12 months ago

I haven't tried with IUPAC characters so far. Let me have a try first.

kokyriakidis commented 12 months ago

I "removed" these characters by randomly substituting them with one of [A, C, T, G]. If you find a better way, it would be wonderful.