Tophat2 alignment problem

[ajsn6c]

I ran this code using tophat (v2.1.0) to align reads (bowtie2 (v2.2.6.0)) from my RNA-seq fastq file using the bowtie2 genomes.bt2 indexes from igenomes (Homo_sapiens_UCSC_hg19)(:

tophat2 -p 8 -G /home/ajsn6c/Desktop/Kumar_RNA-seq/Homo_sapiens_UCSC_hg19/Homo_sapiens/UCSC/hg19/Sequence/Bowtie2Index/hg19.gtf /home/ajsn6c/Desktop/Kumar_RNA-seq/Homo_sapiens_UCSC_hg19/Homo_sapiens/UCSC/hg19/Sequence/Bowtie2Index/genome HPDE_S11_L002_R1_001.fastq

My fastq file is around 13 GB. However, after alignment my accepted hits file is only 50 MB.

Heres the alignment output saying I have around 55 million kept reads:

[2018-02-21 13:58:33] Beginning TopHat run (v2.1.0)
-----------------------------------------------
[2018-02-21 13:58:33] Checking for Bowtie
Bowtie version: 2.2.6.0
[2018-02-21 13:58:33] Checking for Bowtie index files (genome)..
[2018-02-21 13:58:33] Checking for reference FASTA file
[2018-02-21 13:58:33] Generating SAM header for /home/ajsn6c/Desktop/Kumar_RNA-seq/Homo_sapiens_UCSC_hg19/Homo_sapiens/UCSC/hg19/Sequence/Bowtie2Index/genome
[2018-02-21 13:58:35] Reading known junctions from GTF file
[2018-02-21 13:58:39] Preparing reads
left reads: min. length=12, max. length=101, 55970267 kept reads (45104 discarded)
Warning: short reads (<20bp) will make TopHat quite slow and take large amount of memory because they are likely to be mapped in too many places
[2018-02-21 14:17:45] Building transcriptome data files Panc1/tmp/genes
[2018-02-21 14:17:59] Building Bowtie index from genes.fa
[2018-02-21 14:32:14] Mapping left_kept_reads to transcriptome genes with Bowtie2
[2018-02-21 15:38:44] Resuming TopHat pipeline with unmapped reads
[2018-02-21 15:38:44] Mapping left_kept_reads.m2g_um to genome genome with Bowtie2
[2018-02-21 16:17:07] Mapping left_kept_reads.m2g_um_seg1 to genome genome with Bowtie2 (1/4)
[2018-02-21 16:18:13] Mapping left_kept_reads.m2g_um_seg2 to genome genome with Bowtie2 (2/4)
[2018-02-21 16:19:32] Mapping left_kept_reads.m2g_um_seg3 to genome genome with Bowtie2 (3/4)
[2018-02-21 16:20:46] Mapping left_kept_reads.m2g_um_seg4 to genome genome with Bowtie2 (4/4)
[2018-02-21 16:21:59] Searching for junctions via segment mapping
[2018-02-21 16:25:24] Retrieving sequences for splices
[2018-02-21 16:27:18] Indexing splices
Building a SMALL index
[2018-02-21 16:27:37] Mapping left_kept_reads.m2g_um_seg1 to genome segment_juncs with Bowtie2 (1/4)
[2018-02-21 16:27:50] Mapping left_kept_reads.m2g_um_seg2 to genome segment_juncs with Bowtie2 (2/4)
[2018-02-21 16:28:03] Mapping left_kept_reads.m2g_um_seg3 to genome segment_juncs with Bowtie2 (3/4)
[2018-02-21 16:28:17] Mapping left_kept_reads.m2g_um_seg4 to genome segment_juncs with Bowtie2 (4/4)
[2018-02-21 16:28:31] Joining segment hits
[2018-02-21 16:31:02] Reporting output tracks
-----------------------------------------------
[2018-02-22 19:21:42] A summary of the alignment counts can be found in ./tophat_out/align_summary.txt
[2018-02-22 19:21:42] Run complete: 02:08:37 elapse

This is the alignment summary from the align_summary files:

reads:
Input : 926337
Mapped : 898584 (97.0% of input)
of these: 14621 ( 1.6%) have multiple alignments (14 have >20)
97.0% overall read mapping rate.

Why is the input only 900K, when it kept 55 million reads? The quality of the reads have excellent phred scores too. Any ideas would be greatly appreciated!

Thanks
Alex