Question about mapping motif binding sites to genome location

Question

[kanwarjag]

I have a naïve but complex question. I used RSAT to get 5 genes -2000 bp upstream sequence of TSS. I used this FASTA file and binding motif (identified from my experiment) in FIMO to see where is the binding site of the identified motif. I know that protein of interest bound very close to TSS. I get following results from FIMO output:

                                            to get -2000 upstream TSS seq           
# motif_id  motif_alt_id    sequence_name   start   stop    strand  score   p-value q-value matched_sequence    Chr start   End Strand
2       D20_ENSG00000130164-LDLR-ENST00000557958    80  108 +   41.9286 1.14E-14    6.18E-10    CTCTGCCACCCAGGCTGGAGTGCAATGGC   chr19   11102268    11104267    D
2       D102_ENSG00000161048-NAPEPLD-ENST00000425379    106 134 -   37.6735 4.50E-13    1.15E-08    CTCTGTCACCCAGGCTGGAATACAGTGGC   chr7    103128761   103130760   R
2       D17_ENSG00000130164-LDLR-ENST00000558518    309 337 -   32.8367 1.19E-11    1.10E-07    CTCTGTCACCCAGGCTGGAGCGCAGTGAC   chr19   11130163    11132162    D

In the table above column 3 has gene/ transcript name (name Is trimmed from default names because FIMO will not expect long names) column 4 -6 show motif binding regions start, end and strand respectively. Last four columns are Chr, start end and strand corresponding to -2000 bp upstream FASTA that was used as input in FIMO. The problem is I could not figure out how should I location of column 4 and 5 (start and end of motif binding region) to column 12 and 13 that represent original FASTA coordinates.

In row 1: original FASTA (Column14) and motif binding (column 6) are both on forward strand. to get location of column 4 and 5 with in column 12 and 13, should I simply be doing 11102268 +80 and 11104267 – 108. But it does not give me the sequence and insert is also >29. Similarly, if in row 2 both binding motif and FASTA seq are in Reverse strand how should I map to coordinates in FASTA file.

[Satyajeet Khare]

Did you try using BLAT?

P.S. Actually, did you try Ctrl+F in some sequence editor? Sequence editors (e.g. APE) search sequence on either strand. Once you know the pattern, it will be easy in future.

[Malcolm.Cook]

That output is similar to FIMO output format as document here. But not identical. How exactly did you produce it.

Did you use retrieve-seq to get your upstream sequences? What did the first line in that file look like?