Sumstat standardizer fail to handle reverse strand

[hsun3163]

In the snp_match_dup() function, the snp and A0 A1 are handle using this line, which cannot handle the snp reverse issue.

new_query = pd.concat([new_subject.iloc[:,:5],query.loc[pm.qidx].iloc[:,5:]],axis=1)

[hsun3163]

Also the flip should be "TRUE", it is in the allele_flip_qc function()

---

Actually the flipping is not a issue, id of the snp is misleading but the actual ref and alt are correct.

This is due to the order of A0 and A1 in the index are defined by alphabetical order, as shown below. This was kept for getting some other function running. After the dependency are resolved this is to be changed.

def namebyordA0_A1(df,cols=['CHR','POS','A0','A1']):
    df.columns = cols
    prefix = df[[x for x in cols if x not in ['CHR','POS','A0','A1']]+['CHR','POS']].astype(str).agg(':'.join, axis=1)
    names = []
    for p,A0,A1 in zip(prefix,df.A0,df.A1):
        tmp = A0+':'+A1 if A0 > A1 else A1 +':'+ A0
        names.append('_'.join([p,tmp]))
    return names

[hsun3163]

A quick fix to the issue is

, but need to see if there is any residual impact.

[hsun3163]

doesnt seem to be any residual impact, close for now.