Validation methods and systems for sequence variant calls

A computer implemented method for validating variant calls to reduce false positive variant calling from systemic errors, the method comprising:

• receiving a sample containing polynucleotides;
• preparing the sample containing polynucleotides for sequencing;
• sequencing the sample to generate a plurality of sample reads from the polynucleotides;
• wherein said sequencing comprises analyzing multiple different samples in the same sequencing run, thereby reducing sequencing depth of coverage per sample;
• and
• under control of one or more processors executing program instructions for:
• receiving sequencing data in a binary file format, the sequencing data including at least 10,000 sample reads from the plurality of sample reads, each sample read having a corresponding sequence of nucleotides along a genomic sequence of interest;
• receiving an indication of a potential variant call at a designated position within each corresponding sequence of nucleotides along the genomic sequence of interest;
• determining baseline variant frequencies at the designated position within one or more baseline genomic sequences, the baseline variant frequencies representing relative frequencies of a variant at a particular position along one or more baseline genomic sequences in a population, the population corresponding to the number of reads and/or samples obtained from the one or more baseline genomic sequences from a population of normal individuals;
• determining a sample variant frequency at the designated position for the genomic sequence of interest, the sample variant frequency representing a relative frequency of a variant at the particular position along the genomic sequence of interest in a population, the population corresponding to the number of sample reads obtained for the genomic sequence of interest from an individual;
• analyzing the baseline and sample variant frequencies at the designated position to generate a quality score, the analyzing comprising indexing the sample variant frequency with respect to a distribution of the baseline variant frequencies;
• and
• validating the potential variant call for the genomic sequence of interest based on the quality score, the validating comprising comparing the quality score to a threshold using a hierarchical decision tree for multi-level validation, wherein a quality score that exceeds the threshold indicates a valid variant call and a quality score that is less than the threshold indicates a false variant, thereby reducing false positive variant calling from systematic error at a lower depth of coverage;
• wherein the determining of the baseline variant frequencies comprises:
• a) receiving the sequencing data from reference reads for a set of positions within a current base pair window, wherein a length of the current base pair window comprises a plurality of base pairs;
• b) identifying a candidate variant frequency for one or more positions in the set of positions within the current base pair window;
• c) selecting one of the one or more candidate variant frequencies as a single baseline variant frequency for the designated position within the reference reads;
• and
• d) shifting the base pair window along one or more of the baseline genomic sequences and repeating the operations at a), b), and c) to determine baseline variant frequencies at a next designated position.