Entering edit mode
11 weeks ago
eva_u
▴
30
Hi all,
I'm trying to implement t2t assembly for RNAseq alignment with STAR (using updated t2t gtf). First basic run with default STAR settings yielded poor result for t2t - all metrics are lower compared to hg38. Is this because i need to customize STAR parameters to work with t2t? Does anyone have experience doing the same? Curious what your results were.
Thank you!
star script:
#!/bin/bash
...
STAR --runMode alignReads \
--genomeDir ./star_index_hg38/star_index/ \
--readFilesIn ${R1} ${R2} \
--readFilesCommand zcat \
--outSAMtype BAM Unsorted \
--runThreadN 8 \
--outFileNamePrefix STAR.HC2_HH23.hg38_multiMap/${SAMPLE}/
Since one one has commented on this so far I will say that in general T2T genome is going to be of better quality/more complete. Not many use it as default for analysis so you can't go wrong using either. Your plots show the differences to be within a few percentage points of each other, so the differences are not large.
Intuitively it does not seem logical that you would need to use different parameters for a better quality genome. Since your data appears to be RNAseq the regions with genes are likely very similar in both assemblies.