Read length for de bruijn graph

Question

[faraz.k89]

Hi everyone, I am trying to build De Bruijn graph from short reads. I have some reads that has length < 10 (0.01 % only). I am just worried if those reads (however very small % of them) will create problem for graph building?

The stats i am getting for graph building is :

bank                                    
            bank_uri                                 : SRR2847385_interleaved.fasta,SRR2847386_interleaved.fasta
            bank_size                                : 117637467610
            bank_total_nt                            : 89977612926
            sequences                               
                seq_number                               : 455322542
                seq_size_min                             : 1
                seq_size_max                             : 250
                seq_size_mean                            : 197.6
                seq_size_deviation                       : 55.4
            kmers                                   
                kmers_nb_valid                           : 80415655424
                kmers_nb_invalid                         : 3772037
        stats                                   
            histogram                               
                cutoff                                   : 23
                nb_ge_cutoff                       : 332423627
                first_peak                              : 91
            kmers                                   
                solidity_kind                            : sum
                thresholds                               : 3 3 
                kmers_nb_distinct                   : 931511370
                kmers_nb_solid                       : 452752530
                kmers_nb_weak                      : 478758840
                kmers_percent_weak               : 51.4

As you can see large number of them are valid k-mers. Do you think the graph just ignore reads below a certain length?

Thanks in advance. Faraz.

[Sej Modha]

The number of valid reads would depend on the k-mer size specified for the graph.