Sorted bamfiles have increased size!!

[Parashar]

BACKGROUND:
I have RNA-Seq data from Illumina platform.
I aligned the reads post-QC onto hg19 bowtie index using tophat.
Now I wished to get the raw count of reads mapped to each gene using Ht-seq-count.

WHAT I DID:
However the HT-Seq reports error that it is unable to find the mate pair and asks whether the sam file is properly sorted.
Hence, after reading posts on the problem. I sorted my bam files using samtoools.
samtools sort -n accepted_hits.bam accepted_hits_bam_sorted

CONCERN:
Now the sorted BAM file has increased in size and I wish to know why his has happened.
Though it might not be but seems highly unintuitive.

[dpryan]

That's not surprising. Remember that BAM files are compressed, with the compression algorithm being more efficient the more similar neighbouring reads are (i.e., coordinate-sorted files should compress to a greater extent).

[Parashar]

Thanks dpryan.
Just wanted to confirm if something went horribly wrong.
I'll now be following your earlier reply on a different post to run HTSeq

[maubp]

You sorted by read name (-n), but the default sorting by mapping position puts similar reads together, and therefore their sequence data compresses well, so coordinate sorted BAM files are usually smaller.