Andrew T. Martens

---

This is a summary of how to use the code to perform the data analysis on the Linux (BASH) command line. Original work was performed on a computer running Fedora 20 Linux in 2013-2015. It helps to put the perl programs, genomes files, and so on in their own directories, but is not required.

1. Get a Sequence Read Archive (SRA) file from the Gene Expression Omnibus. Example: SRR1067765.sra

2. Convert the SRA files to FASTQ files

fastq-dump ./SRR1067765.sra -O .

3. Trim the ends

fastx_clipper -Q33 -a CTGTAGGCACCATCAAT -l 21 -c -n -v <SRRxxx.fq >SRRxxx.trimmed.fq

4. Build the Bowtie rRNA/tRNA database

bowtie-build -f rrna-seqs.fa rrna_seqs

5. Concatenate the trimmed FQ files

cat SRR1067765/SRR1067765.trimmed.fq SRR1067766/SRR1067766.trimmed.fq SRR1067767/SRR1067767.trimmed.fq SRR1067768/SRR1067768.trimmed.fq > redo_calcs/Sequencing_Reads/trimmed-data.fq

6. Use rRNA/tRNA database to filter out these sequences with Bowtie 1.0

bowtie --quiet -p 8 -l 23 --un=Sequencing_Reads/norrna.fq rrna_seqs -q Sequencing_Reads/trimmed-data.fq > /dev/null

7. Align to genome. This means allow up to 2 mismatches in the first 23 nucleotides, no multi-mappers. Whether or not to allow multi mappers depends on the tradeoff: which is worse - to assign read density where it doesn't belong, or to remove read density from where it does? When performing global analyses of the data, perhaps disallowing multi-mappers is best, but when doing gene-specific analyses, perhaps it's best to leave them. There is no correct answer.

bowtie -S -p 8 -l 23 -m 1 --sam-nohead e_coli_MG1655 -q norrna.fq > aligned.SAM

Result:

reads processed: 158173518

reads with at least one reported alignment: 95854989 (60.60%)

reads that failed to align: 54522442 (34.47%)

reads with alignments suppressed due to -m: 7796087 (4.93%)

Reported 95854989 alignments to 1 output stream(s)

8. Only keep the parts that matched the genome. (0 for unmapped, 1 for multi mapped)

grep -v -E 'XM:i:[01]' aligned.SAM > matched.SAM

Left with 95854989 reads

9. Make a "simplified" SAM file.

Note: does not properly account for reads which are the "wrong" direction (these events are rare, probably due to trace DNA contaminants).

./Perl_Programs/fix_SAM_multi_chromosome.pl Genomes/ecoli_chromosome/ Sequencing_Reads/matched.SAM Sequencing_Reads/fps.simple

10. Process the PTT file to remove any ORFs which have in-frame stop codons.

./Perl_Programs/ptt_no_stop_multi_chromosome.pl Genomes/ecoli_chromosome/ Genomes/ecoli_ptt/ Genomes/ecoli_ptt_nostops/

11. Filter out any sequences which are not "elongation" sequences.

a. Must be in protein coding sequences, in the PTT file (with no in frame stops)

b. Must be within 10 codons from the ends

Note: this runs into problems with overlapping ORFs, which fortunately are uncommon.

./Perl_Programs/filter_only_elong_footprints_multi_chromosome.pl Genomes/ecoli_ptt_nostops Sequencing_Reads/fps.simple Sequencing_Reads/elong.simple

12. Make heatmaps of extended footprint codon composition by length class. Frame matters. Can either be aligned to the 5' end (LtoR) or the 3' end (RtoL), and may or may not be normalized to mRNAseq data. Can choose to fill in sequence from the genome, as well. Example (forgive the terrible file naming scheme):

./Perl_Programs/footprint_codon_composition_lengths_position_simple_heatmaps_RtoL_multi_chromosome.pl Genomes/ecoli_chromosome/ Genomes/ecoli_ptt_nostops/ Sequencing_Reads/elong.simple Plots/codon_heatmaps/ Plots/codon_heatmaps/len_tally.tsv

13. Extend the 5' ends to 45 nt, discard anything > 45 nt. Frame does not matter.

./Perl_Programs/footprint_extend_5prime_track_length_multi_chromosome.pl Genomes/ecoli_chromosome/ Genomes/ecoli_ptt_nostops/ Sequencing_Reads/elong.simple Sequencing_Reads/extended.simple

14. Make GC heatmap plot.

./Perl_Programs/extended_fps_gc_content_v2.pl Sequencing_Reads/extended.simple Plots/gc_heatmap

15. Make 2d-heatmaps of footprint ends

a. Find the endpoints and append to files (this means must manually delete output from a previous run)

./Perl_Programs/footprint_2d_heatmap-simple.pl Genomes/ecoli_chromosome/NC_000913.fna Genomes/ecoli_ptt_nostops/NC_000913.ptt Sequencing_Reads/elong.simple Plots/2d_heatmaps/endpoints/

b. Tally the endpoints and arrange into 2d matrix

./Perl_Programs/footprint_2d_heatmap_part2_nucleotide_multi_chromosome.pl Genomes/ecoli_ptt_nostops/NC_000913.ptt Plots/2d_heatmaps/endpoints/ Plots/2d_heatmaps/out/

Then, need to go manually in Gnuplot and make the plot with the appropriate x,y range etc. example: set datafile missing "-" set key off plot "rpsA-b0911.heat.txt" matrix with image