search

KennthShang / CHERRY

Host prediction for phages
GNU General Public License v3.0
20 stars 8 forks source link

multimodal Graph Error for file contig_0 #6

Open dangchenyuan opened 1 year ago

dangchenyuan commented 1 year ago

I want to use Cherry to predict the host with my own MAGs. I replaced the file prokaryote and prokaryote.csv, and run the command "python run_Speed_up.py --contigs final_quality_summary.fasta --len 5000 --model pretrain --topk 1" . Unfortunately, I met this error as follows:

image

KennthShang commented 1 year ago

Hi there,

Can you show what information you have added to the "prokaryote.csv" file?

BTW, the "prokaryote.csv" should not be replaced, but new information add to it.

Also, please follow the naming rules introduced in the guidelines. It seems your file is named as "4.26An_bin.37.strict.fa". Please do not use so much '.' in the name to avoid potential problems (it seems the BLAST identified your file name as '4' shown in the second line of your screenshot). For example, you can rename it as "4_26An_bin_37_strict.fa" and add information for accession "4_26An_bin_37_strict" in the "prokaryote.csv" file.

Best, Jiayu

dangchenyuan commented 1 year ago

Thanks. I know what the problem is.
I'm also wondering when I use my own MAGs whether the CRISPR spacer match method proceeded during the prediction process.

dangchenyuan commented 1 year ago

Hi Jiayu, I re-ran my command as following your comments, but still encountered an error. I compared this false MAG and other worked MAG but found no difference.

image

KennthShang commented 1 year ago

Hi there,

For your first question: Yes, it will use the CRISPR information. However, this information is from the CRISPR database. NOT the CRISPR from your MAGs. If you want to predict the CRISPR between your MAGs and your phages, please try the lite version introduced in the guidelines (I also attached the link here for your quick check https://github.com/KennthShang/CHERRY_crispr). This lite program is relatively fast compared to the main program.

The second question: The error message here is from the alignment results against the CRISPR database. As I said, you cannot just REPLACE the "prokaryote.csv" but add your MAG information at the end of the original "prokaryote.csv". Otherwise, the CRISPR process will counter this bug.

Best, Jiayu