CD-HIT can not complete remove redundant transcript

Question

[xioli2013]

I tried to use CD-HIT-EST to remove redundancy in trinity de novo transcripts however I can still see redundant annotations,

cd-hit-est -i trinity.fasta -o clstr_out -c 0.9 -n 9

for example:

TRINITY_DN1855_c5_g1, TRINITY_DN1855_c1_g1

all pointing to dnaK, the two sequences aligned at 92% identity but they are not clustered by CD-HIT

>TRINITY_DN1855_c5_g1 
CGCCAAGAAGACCGAGATCTACAGCACCGCCGAAAACAACCAGCCCGGTGTGGAAATCAACGTGCTGCAAGGCAAGCGCC
CCATGGCCGCCGACAACAGGTCCCTGGGCCGCTTCAAGCTCGAGGGCATTCCCCCCATGCCCGCAGGCTGCGCCCAGATC
GAAGTGACCTTCGGTATCGACGCCAACGGCATTCTGCATGTCACCGCCAAGGAAAAGACCAGCAGCAAGGAAAGCAGCAT
CCGCATCGGGAACACCACCACCCTCGACAAGAGTGACGTGGAGCGCATGGTGCAGGAAACCGAGCAGAACGCCGCCGCCG
ACAGGGCCCGCAAGGAGAAGGTCGAGAAACGCAACAACCTCGACTCGCTGCGC
> TRINITY_DN1855_c1_g1
AGGGCGGCATGATTGCCCCGATGGTTACCCGCAACACCACCGTGCCCGTCAAGAAGACCGAGATCTACACCACTGCCGAAAA
CAACCAGCCCGGCGTGAAAATCAACGTGCTGCAAGGCGAGCACCCCATGGCCGCCGACAACAAGTCTCTGGGCCGCTTCAAGCTCGAAGGCGTTCCCCCCATGCCCGCAGGCCGCGTCCAGATCGAAGTGACCTTCGATAT

Trying other parameters as -c 0.89, 0.88 did not reduce the redundancy but actually increased the number of transcripts.

I am writing to hear your comments as to what the problem is and how to address the issue

Thanks,

Xp

[GenoMax]

See this thread for further assistance: how to use CD_HIT to remove the redundant sequence from trinity output file

Answer 1

[lakhujanivijay]

There are inherent limitations to CD-HIT algorithm which we should be aware of. Please see this link

Also, see CD-HIT-2D comparing algorithm on the same page