Reads should be mapped to two genes but mapped to one instead using STAR.

[Jin Li]

Hi all,

I have a paired end reads that can be mapped perfectly to the gene Lsm7 and Gm10146, but STAR output the alignment of Lsm7 only.

The paired end reads in use are as below.

The first read in pair:

```

@HWI-ST1280:274:C4MA1ACXX:1:1103:12963:29417/1

GTCCTGCTGCTGCACGAAGGGGTTGGGGATGGCCTCCATGCCGTCCTGTGGGCAGATGAGCACCACCGAGGTGCCACGACATACGACAAGCCCCAGCTGC

+

CCCFFFFFHHHHHJJJJJJJJJHHJJJIIIJJJIJJJJJJIJJIJJJJHHHHFFFDEEEEEDDDDDDDDDD;BDDCCBDDDDDDDDDDDDBDDDBBDDDD

```

The second read in pair:

```

@HWI-ST1280:274:C4MA1ACXX:1:1103:12963:29417/2

AGGGCTTTGACCCACTGCTCAACCTGGTGCTGGATGGGACCATGGAGTACATGAGAGACCCTGATGACCAGTACAAGCTGACGGAGGACACTCGGCAGCT

+

CCCFFFFFHHHHHJJJJJJJJJJJJJJFHIJJJJIJJJJJJIJJJJIFHIFHIJJJJJJJJJJJJJIHHHHHFFFFFFEEEEDDDDDDDDDDDDDDDDDC

```

The perfect alignments were obtained using BLAT in UCSC genome browser website.

```

   ACTIONS      QUERY           SCORE START  END QSIZE IDENTITY CHRO STRAND  START    END      SPAN
---------------------------------------------------------------------------------------------------
browser details HWI-ST1280:274:C4MA1ACXX:1:1103:12963:29417_1   100     1   100   100 100.0%    10   -   78393526  78393625    100
browser details HWI-ST1280:274:C4MA1ACXX:1:1103:12963:29417_1   100     1   100   100 100.0%    10   +   80852964  80853063    100

browser details HWI-ST1280:274:C4MA1ACXX:1:1103:12963:29417_2   100     1   100   100 100.0%    10   +   78393432  78393531    100
browser details HWI-ST1280:274:C4MA1ACXX:1:1103:12963:29417_2    99     1   100   100 100.0%    10   -   80853058  80853814    757

```

The perfect alignments of the paired-end reads are mapped to Gm10146 (first read: chr10: 78393526-78393625, second read: chr10: 78393432-78393531) and Lsm7 (first read: chr10: 80852964-80853063, second read: chr10: 80853058-80853814).

However, STAR will return the alignment that was mapped to Lsm7 only as the following.

```

HWI-ST1280:274:C4MA1ACXX:1:1103:12963:29417 99 chr10 80852964 255 100M = 80853058 851 GTCCTGCTGCTGCACGAAGGGGTTGGGGATGGCCTCCATGCCGTCCTGTGGGCAGATGAGCACCACCGAGGTGCCACGACATACGACAAGCCCCAGCTGC CCCFFFFFHHHHHJJJJJJJJJHHJJJIIIJJJIJJJJJJIJJIJJJJHHHHFFFDEEEEEDDDDDDDDDD;BDDCCBDDDDDDDDDDDDBDDDBBDDDD NH:i:1 HI:i:1 AS:i:200 nM:i:0

HWI-ST1280:274:C4MA1ACXX:1:1103:12963:29417 147 chr10 80853058 255 43M657N57M = 80852964 -851 AGCTGCCGAGTGTCCTCCGTCAGCTTGTACTGGTCATCAGGGTCTCTCATGTACTCCATGGTCCCATCCAGCACCAGGTTGAGCAGTGGGTCAAAGCCCT CDDDDDDDDDDDDDDDDDEEEEFFFFFFHHHHHIJJJJJJJJJJJJJIHFIHFIJJJJIJJJJJJIJJJJIHFJJJJJJJJJJJJJJHHHHHFFFFFCCC NH:i:1 HI:i:1 AS:i:200 nM:i:0

```

Please also find the attached `Log.out` from STAR.

Could you help clarify why STAR outputs only one perfect alignment instead of both of the two perfect alignments? Thank you.

Best regards,

Jin

[Alexander Dobin]

Hi Jin,

by default, STAR favors annotated spliced alignments over unspliced (or unannotated spliced).

This is controlled by --sjdbScore = 2 by default. If you reduce this parameter to 0 (or even 1), you should see both alignments.

Cheers

Alex