htseq read counts different in separate runs

[h_manoj]

Hello,

I had used tophat2 to map a set of fastq files. I noticed that the reads mapped for some features were different in two separate runs (for eg., ENSMUSG00000047905 produced 4335 reads in one run while it shows only 2 reads in another run), even though I had kept all options constant in both runs. The only change was the annotation file. In the first try, I used the gencode M2 gtf file which contained only level 1 and level 2 features. In the second trial, I had used the complete gtf file. But, the feature (ENSMUSG00000047905) that I am checking has no difference in either annotation files.

Details of ENSMUSG00000047905 in gencode level 1 and level 2 annotation file:
login3$ grep ENSMUSG00000047905 gencode.vM2.annotation.gtf |grep -w gene
chr6 HAVANA gene 67205619 67206083 . + . gene_id "ENSMUSG00000047905.10"; transcript_id "ENSMUSG00000047905.10"; gene_type "pseudogene"; gene_status "KNOWN"; gene_name "Gm8566"; transcript_type "pseudogene"; transcript_status "KNOWN"; transcript_name "Gm8566"; level 2; havana_gene "OTTMUSG00000024710.1";

Details of ENSMUSG00000047905 in gencode complete annotation file:
login3$ grep ENSMUSG00000047905 gencM2_levels1and2.gtf |grep -w gene
chr6 HAVANA gene 67205619 67206083 . + . gene_id "ENSMUSG00000047905.10"; transcript_id "ENSMUSG00000047905.10"; gene_type "pseudogene"; gene_status "KNOWN"; gene_name "Gm8566"; transcript_type "pseudogene"; transcript_status "KNOWN"; transcript_name "Gm8566"; level 2; havana_gene "OTTMUSG00000024710.1";
login3$

What would be the reason for this? Is this expected? Appreciate any help!
Thanks.