FASTQ vs BAM

[RViuj]

Hi,

I tried with a paired end RNA Seq FASTQ and even after almost 8-10 hours, the sample run did not result in any outputs. Instead after reading 48.7 M reads, it did not progress. I wonder why ? I used 16 CPU threads and also tried with -k 5 & 7 "~/TRUST4/trust4 -f ~/TRUST4/reference/human_IMGT+C.fa
-1 ~/Desktop/dtFASTQ/d001_R1.fastq.gz
-2 ~/Desktop/dtFASTQ/d001_R2.fastq.gz
-o ~/Desktop/dtFASTQ/TRUST4_results/D001
-t 16
-k 7

I decided to use bam files for the next try. Could you please suggest if the BAM files need to be sorted and filtered before running TRUST4 or can I just use the bam files as input without any sorting and filtering ?

Your expert reply will be very help. Thank you.

[mourisl]

What did TRUST4 output on the screen? Is your data bulk RNA-seq data or targeted amplified TCR/BCR data? For BAM input, it needs to be sorted by coordinate. No need for filtering.

[RViuj]

Thank you. After "Read in and count kmers for 44800000 reads." although the CPU usage was around 350%, it did not progress. There were no error messages or TRUST4 was not stuck (ps aux | grep trust4) nor no output folder/files (even empty) were created. There was no progress even after 4-5 hours and it just remained blank.

Thank you for the information about BAM file preparation.

[mourisl]

The step that stalls is sorting all the candidate reads, which can be memory intensive given you have 45M reads. If you are running on your local machine, and if it runs out of memory, the speed will be very slow.

[RViuj]

Thank you. I created the BAM files to run TRUST4. However,

1 - TRUST4 did not recognise the BAM files. To make sure the BAM file is correct:
a - Checked the BAM file and confirmed it was valid using samtools quickcheck.
b - Ensured that the BAM file was indexed correctly with the samtools index command.
c - Verified that the BAM file had read groups (@rg).
d - Despite all these checks, TRUST4 kept returning the help menu.

Then I checked it the TRUST4 was complied with BAM Support, I checked for HTSlib (which I thought is required for BAM file processing).

a - TRUST4 returned no link to HTSlib
b - I modified the Makefile to explicitly link TRUST4 with HTSlib using -L$(brew --prefix htslib)/lib -lhts.
c - Tried forcing HTSlib during compilation with specific paths.
d - Despite these changes, otool (which shows the linked libraries) never showed libhts.dylib, indicating BAM support was not enabled during the build process.
e - I recompiled TRUST4 multiple times, but BAM support was still missing.

Then, I tried to fix TRUST4 to include necessary libraries (like libhts.dylib), preventing BAM support from being enabled.
a - I installed HTSlib using Homebrew and attempted to recompile TRUST4 with the correct HTSlib paths.
b - Checked for errors and attempted to link HTSlib manually.
c - Despite numerous attempts, HTSlib was never properly linked in the TRUST4 binary.

After all, reinstalled TRUST4 via CONDA. Even then TRUST4 failed to recognise BAM input.

While TRUST4 was recognising FASTQ files, due to the storage/memory, it never finished the run. When the BAM file was used as the input, it is not recognising. I am not sure how I can fix it. Your expert reply will be helpful. Thank you.

[RViuj]

I also figured that "libbam" file is not built when samtools is installed either from the sourcecode (https://github.com/samtools/samtools/releases) or using Homebrew. May I know how I could make this to run TRUST4 using BAM as input? Thank you.