How should the fl_count matrix be formatted? Using output from IsoQuant.

[andysaiz]

I am trying to run SQANTI3 QC and it processes through much of the pipeline but fails every time when it gets to the fl_count matrix due to this error. It is detecting an accepted counts matrix format but fails afterwards.

I am supplying the IsoQuant output file called: OUT.discovered_transcript_grouped_counts.tsv and I have tried modifying it a few different ways.

[INFO:2025-07-28 21:17:46,733] Multi-sample PacBio FL count format detected.
Traceback (most recent call last):
  File "/mnt/data/sqanti3/release_sqanti3/sqanti3_qc.py", line 52, in <module>
    main()
  File "/mnt/data/sqanti3/release_sqanti3/sqanti3_qc.py", line 41, in main
    combine_split_runs(args,split_dirs)
  File "/mnt/data/sqanti3/release_sqanti3/src/parallel.py", line 197, in combine_split_runs
    write_classification_output(isoforms_info, outputClassPath, FIELDS_CLASS)
  File "/mnt/data/sqanti3/release_sqanti3/src/qc_output.py", line 42, in write_classification_output
    fout_class.writerow(isoforms_info[iso_key].as_dict())
  File "/mnt/data/miniforge3/envs/sqanti3/lib/python3.11/csv.py", line 154, in writerow
    return self.writer.writerow(self._dict_to_list(rowdict))
                                ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/mnt/data/miniforge3/envs/sqanti3/lib/python3.11/csv.py", line 149, in _dict_to_list
    raise ValueError("dict contains fields not in fieldnames: "
ValueError: dict contains fields not in fieldnames: 'FL.NT_rep2', 'FL.NT_rep3', 'FL.STRAP_clone_1_rep2', 'FL.STRAP_clone_2_rep2', 'FL.STRAP_clone_2_rep3', 'FL.NT_rep1', 'FL.STRAP_clone_1_rep3', 'FL.STRAP_clone_2_rep1', 'FL.STRAP_clone_1_rep1'
ERROR: There was an unexpected issue during SQANTI3 QC module execution

The default output format of the counts matrix from IsoQuant is:

gene_id	NT_rep1	NT_rep2	NT_rep3	STRAP_clone_1_rep1	STRAP_clone_1_rep2	STRAP_clone_1_rep3	STRAP_clone_2_rep1	STRAP_clone_2_rep2STRAP_clone_2_rep3
transcript1.chr1.nnic	0	0	0	0	0	0	0	1.0	1.0
transcript1.chr10.nnic	0	0	1.0	1.0	0	1.0	0	0	0
transcript1.chr14_GL000194v1_random.nnic	0	0	0	0	1.0	1.0	0	0	0

I tried modifying it a few different ways to match the SQANTI3 documentation
Version 1:

superPBID	NT_rep1	NT_rep2	NT_rep3	STRAP_clone_1_rep1	STRAP_clone_1_rep2	STRAP_clone_1_rep3	STRAP_clone_2_rep1	STRAP_clone_2_rep2	STRAP_clone_2_rep3
transcript1.chr1.nnic	0	0	0	0	0	0	0	1.0	1.0
transcript1.chr10.nnic	0	0	1.0	1.0	0	1.0	0	0	0

Version2:

id,sample1,sample2,sample3,sample4,sample5,sample6,sample7,sample8,sample9
transcript1.chr1.nnic,0.0,0.0,0.0,0.0,0.0,0.0,0.0,1.0,1.0
transcript1.chr10.nnic,0.0,0.0,1.0,1.0,0.0,1.0,0.0,0.0,0.0
transcript1.chr14_GL000194v1_random.nnic,0.0,0.0,0.0,0.0,1.0,1.0,0.0,0.0,0.0

The error was generated when I ran it using Version 1. I did not use PacBio IsoSeq, so the transcript IDs are different (e.g. transcript1.chr1.nnic), but I figured this would be okay. How should the fl_counts matrix be formatted or are there any other fixes? I am out of ideas!

Anyone run SQANTI3 with the files from the IsoQuant output?

[andysaiz]

I was able to solve this by replacing a line within sqanti3/release_sqanti3/src/parallel.py to allow handling of different sample names:

Within parallel.py:

#TOOK OUT THE COMMENTED LINE and replaced it with the line below to fix problem of not recognizing sample-specific columns: #from fl_count matrix. #write_classification_output(isoforms_info, outputClassPath, FIELDS_CLASS) write_classification_output(isoforms_info, outputClassPath, fields_class_cur)