Pandora index failed on prg output

[mthang]

I am experencing the following error while running pandora index and unable to find the solution. I suspected it was lack of RAM at the beginning, but the error still appears after increase the RAM size to 24G. The input prg fasta file contains 15625 and I have little idea how much memory is required. After having a closer look at my input file, I realized the sequence(s) causing this issue contains numbers. I would like to know how Pandora index handles these sequences.

ERROR
terminate called after throwing an instance of 'FatalRuntimeError'
what(): [FATAL ERROR]: In conversion from linear localPRG string to graph, splitting the string by the next var site resulted in the wrong number of intervals. Please check that site numbers are flanked by a space on either side. Or perhaps ordering of numbers in GFA is irregular?! Size of partition based on site 5 is 1

[iqbal-lab]

Can you give a bit more detail. What exactly was the file it failed on- would it be possible to share it, on this issue if small or via some other means. Also, I'm afraid we are on vacation now until 6th January. Will reply ASAP afterwards

[mthang]

@iqbal-lab Thank you for speedy response ! I didnt expect to receive the response so quick. I understand it's a holiday season, so this issue can wait. I ran the pandora index on a file which was generated by make_prgs program with parameter from_msa. What I understand is the program make_prgs is used to generate a graph base fasta from multiple alignment file. What interesting is the output of make_prgs contains numbers in the sequences. I expect Pandora index woudl work with graph base sequences file. This is also mentioned in this post https://cpang.netlify.app/post/testing-pandora-core/ . It would be great if you can have a look and let me know how to resolve this. The pandora program I used is the binary version (0.12.0). Here's the file I used
Uploading prg.zip…

[iqbal-lab]

The output of make_prg does contain numbers. The format , I now realise, is not described on the github and we will fix that. Bit if you look at Figure 2 here

https://www.biorxiv.org/content/10.1101/059170v2.full

you'll get the idea. Will be in touch in the new year!