Missing cbai transcriptome sequences?

[sr320]

Looking at this notebook post we generalized

cbai_transcriptome_v2.0.fasta

could be separated into

cbai_transcriptome_v2.1.fasta. + hemat_transcriptome_v2.1.fasta

However number of sequences is

cbai_transcriptome_v2.0.fasta (1.4M)

could be separated into

cbai_transcriptome_v2.1.fasta (237k) + hemat_transcriptome_v2.1.fasta (30k)

Q: should there be this big a "loss" of sequences given the approach?