Fixing FASTA file for Local BLAST Database

Question

I recently prefetched 157 SRA files from an NCBI BioProject using the SRAtoolkit. I then used the toolkit to download those files in FASTA format.

Each individual FASTA file looks something like this:

(example for SRR5678544 file)

>SRR5678544.1 HWI-ST1146_88:5:1112:6472:81473 length=194
CCATGCGGGGTATCGTATGCTTCCTTCTGCACTACCCTTTAGCTGTTCTATATGCTGCCACTCCTCAATTGGATTAGTCTCATCCTTCAATGCTATCACAAGAAGTAGAGACACAAATGCAAGAGGAGCATATAAATTACAAAACACCATCACTGAGGGCCCTAAAGCGGTTCCCACGAAAAAAAGGAGAGTAG   
>SRR5678544.2 HWI-ST1146_88:5:1113:13218:62635 length=194
 TGGATGGTGGTTGGGAGTGGTAAGGTTGAATGAGACACGGTAACGAGTGGGGAGGTAGGGTAATGGAGGGTAAGTTGAGAGACAGGTTCGTCAGGGGACACACCACACCACACCCACACCCACACACCCACACACCACACCCACACACCACACCCACACCCACACCCCCACACACCACACCCCCACCCCACCCA  
>SRR5678544.3 HWI-ST1146_88:5:2206:11224:22269 length=194
 GTCTCTTAACTTACCCTCCATTACCCTACCTCCCGACTCGTTACCCTGTCTCATTCAACCATACCACTCCCGACCACCATCCATCCCTCTACTTACTACAGTATGGTGAGTGGGACATGGTGGATGGTAGGGTAAGTGGCAGTGGAGTTGGATATGGGCAATTGGAGGGTAACGGTTATGGTGGACGGGGGGTG

And so on and so forth. (Ignore the " and ' marks I had to input those to format it right on here)

Question 1: Can someone explain why this file has so many sequences (>)? Every FASTA file I have worked with has one line with ">" at the top and a header sentence followed by a single big paragraph of the genome sequence.

Question 2: If I wanted to combine all the sequences in a single FASTA file to look like a "normal" FASTA file - just a single paragraph - how would I do that?

Reasoning: I'm asking simply because I need to make a BLAST database out of the 157 sequences and the "makeblastdb" command only takes in a single file as the input. I was going to try and combine each individual FASTA file into single paragraphs and then make a huge FASTA file that includes all 157 sequences separated by ">".

Answer 1

Briefly the best strategy I think is to download fastQ files and process this via an assembler. I wouldn't enter raw short read Illumina data into a local Blast database. What I'm pretty certain of is thats a bad move.

Which assembler I personally think is a separate question on Bioinformatics SE. It will depend on the genome, for large eukaryote genomes BWA-mem2 seems popular, but its not my thing and there's other stuff upstream and downstream (someone might dispute that). For assembling small genomes, Spades is popular, but preferences are changing and there's also upstream QC. The other question is whether you have a reference genome to assist assembly.

---

The update is this a yeast genome with a reference. This is complicated because its a small genome but will comprise heterozygosity. If it was me - i.e. someone way outside the yeast community - I'd use Spades and provide a reference genome to perform an assembly. However, I'm not sure thats at all cool, because you will will reduce the eukaryotic genome to a haplotype. This could get complicated because what you would be left with may not be biologically real, e.g. two alleles at different parts of the same chromosome on different strands may result in a in silico artefact. I think its a separate question personally.